2026-10-23
Career & Learning Craft
Lesson
Mental Model Transfer: Moving a Model Into a New Domain
Not a picture. A map of a structure. Parts plus their relations. Invariants that never break. Queue: a line, one served at a time.
machine-learning
2026-10-22
AI/ML
Lesson
Semi-Supervised Learning: When Unlabeled Data Helps, and When It Backfires
A few labels anchor the geometry. The unlabeled mass supplies the shape.
machine-learning
2026-10-22
Cybersecurity Career Mastery
Lesson
Security Operations Centers: Turning Telemetry into Detection and Triage
A Security Operations Center (SOC) turns telemetry into decisions through an eight-stage chain: collect → normalize → enrich → detect → triage → investigate → contain → learn . The job is not to fire
machine-learning
2026-10-21
Cybersecurity Career Mastery
Lesson
Cybersecurity Career Paths: Choosing Blue Team, Red Team, AppSec, Cloud, or GRC
Security roles differ mainly by the feedback loop they own . Blue-team and security-operations roles detect and contain live behavior . Red teams test assumptions under authorization . Application sec
machine-learning
2026-10-21
Career & Learning Craft
Lesson
Capture Is Storage, Connection Is Recall
machine-learning
2026-10-20
Cybersecurity Career Mastery
Lesson
Security-Relevant Systems: Processes, Permissions, Packets, and Trust Boundaries
Run as a dedicated account. The principal can no longer reach the host, so the blast radius shrinks to the service.
machine-learning
2026-10-20
Security & Identity
Lesson
Certificate Chains: Why Your Browser Trusts a Site
machine-learning
2026-10-19
Search Systems & Retrieval
Lesson
Search Modes: Choosing Exact, Full-Text, Semantic, Fuzzy, and Geospatial Retrieval
machine-learning
2026-10-19
AI/ML
Lesson
Greedy vs Random: How the Next Token Is Chosen From a Probability Distribution
machine-learning
2026-10-18
Observability & Reliability
Lesson
Log Aggregation: From Distributed Events to Searchable Evidence
Applications, hosts, containers, network devices, and managed services emit events. A durable buffer absorbs bursts before parsing can keep up. Parsers turn heterogeneous lines into fields — timestamp
machine-learning
2026-10-18
AI/ML
Lesson
AdamW: How Momentum and Per-Parameter Learning Rates Steer LLM Training
Panel 1 (full-width horizontal flow diagram, five boxes with arrows, short labels): BATCH -> MODEL FORWARD (tokens to a loss number) -> LOSS (one number) -> BACKPROP (chain rule: one gradient per weig
machine-learning
2026-10-17
Retrieval & Vector Search
Lesson
HNSW vs IVF: Graph Traversal, Cluster Probing, and the Recall-Latency Trade-off
Exact nearest-neighbour search scores a query against every vector — honest, but it does not scale. HNSW and IVF are two different ways to make the search approximate: HNSW is a proximity graph you wa
machine-learning
2026-10-17
AI/ML
Lesson
Flash Attention: Why the NxN Matrix Is a Memory Problem, Not a Compute One
S1 The bus, not the math: Tensor Cores only compute on data in on-chip SRAM; HBM is slow off-chip memory. The NxN attention score matrix is too big for SRAM, so it lives in HBM. The bottleneck is memo
machine-learning
2026-10-16
AI/ML
Lesson
Pre-Norm vs Post-Norm: How a Clean Residual Highway Lets GPT Train 96 Layers
machine-learning
2026-10-16
AI Agent Engineering & Reliability
Lesson
Agent Workflow Ordering: Launch Long-Running Chains Before Side Tasks
The longest independent operation. It defines the critical path — if it never starts, nothing else you did matters for completion.
machine-learning
2026-10-15
Career & Learning Craft
Lesson
Structured Technical Interviews: Rubrics and Panel Calibration Over Gut Feel
Mental model: a panel is a measuring instrument . The rubric is its scale, calibration is checking the scale on a known sample, and independent scores are the reason the instrument is worth running at
machine-learning
2026-10-15
AI/ML
Lesson
Causal Mask vs Padding Mask: The Two Attention Gates, and When Each Fires
A mask is a bias in. -inf before softmax. Zero weight out.
machine-learning
2026-10-14
AI/ML
Lesson
Grouped Query Attention: Why Every Query Head No Longer Needs Its Own KV Cache
Cache memory tracks KV heads, never query heads.
machine-learning
2026-10-14
Caching & Data Freshness
Lesson
Cache-Aside vs Write-Through: Who Owns Cache Population
Cache population is not a single policy — it is two separate paths, and each path can be owned by a different actor.
machine-learning
2026-10-13
NLP Foundations & Text Representation
Lesson
Stemming, Lemmatization, and Stopword Removal: When Text Preprocessing Helps and When It Hurts
Crude suffix-stripping rules (Porter, Snowball): running → run , studies → studi . Fast, rule-based, and the output need not be a real word — that is accepted, because the goal is token collision, not
machine-learning
2026-10-13
AI/ML
Lesson
Encoders and Decoders: Where Q, K, and V Come From
The table row is input. Attention makes the context.
machine-learning
2026-10-12
Web Performance & Observability
Lesson
The Server-Timing Header: Splitting TTFB into Database, Cache, and App Time
The browser's network panel already splits a request into DNS lookup, connection, request sent, Waiting (TTFB), and content download . But everything the server did to produce the response — queries,
machine-learning
2026-10-12
AI/ML
Lesson
Semantic Embeddings: How Context Proximity Shapes Vector Space
Meaning is geometry: co-occurrence trains the distance, and one word can hold one vector — or a fresh one for every sentence it meets.
machine-learning
2026-10-11
System Design
Lesson
The Dependency Rule: Clean vs Onion Architecture
One arrow, two schools — and a one-command test that exposes the fakes.
machine-learning
2026-10-11
Web Security & Browser Isolation
Lesson
Reverse Tabnabbing: How window.opener Lets a New Tab Redirect the Old One
When you open a link with target="_blank" , the browser hands the new page a window.opener reference back to the page that launched it. Same-origin policy still blocks the new page from reading your D
machine-learning
2026-10-10
System Design
Lesson
How to Snapshot a Running Distributed System
You cannot stop the world to photograph it, yet you need a global state that never shows an effect before its cause . The Chandy–Lamport snapshot takes a CUT: one local state per process, plus one cha
machine-learning
2026-10-10
Algorithms & Data Structures
Lesson
Comparison Sorts: Why the Lower Bound Is n log n
A comparison sort learns about the input only by asking pairwise questions — “is a less than b?” — and never looks inside the keys. Model any deterministic comparison sort as a binary decision tree :
machine-learning
2026-10-09
System Design
Lesson
The Kinesis Data Streams architecture
One log per shard, one cursor per reader.
machine-learning
2026-10-09
Cloud Security & Vulnerability Management
Lesson
Amazon Inspector: Prioritizing Findings Across EC2, ECR, and Lambda
Amazon Inspector is an automated vulnerability-management service on AWS. Once enabled, it continuously evaluates supported resources — EC2 instances, ECR container images, and Lambda functions — and
machine-learning
2026-10-08
System Design
Lesson
Web Server vs. Application Server: Core roles
Same hardware, two opposite jobs. The split is about workload, not machine.
machine-learning
2026-10-08
Distributed Systems
Lesson
A System That Is Always Up and Frequently Wrong
Did it answer at all? Measured as successful responses ÷ total requests — the uptime fraction.
machine-learning
2026-10-07
AI/ML
Lesson
What is a RAG feedback loop?
User signals sharpen the ranker — if read honestly.
machine-learning
2026-10-07
NLP Evaluation & Reliability
Lesson
BLEU, ROUGE, and Perplexity: What Each Metric Actually Measures
n-gram precision against references + a brevity penalty . Built for corpus-level machine-translation comparison.
machine-learning
2026-10-06
AI/ML
Lesson
LLM Cascades: Escalate Only When the Cheap Model Fails
Cheap model first. The check decides the rest.
machine-learning
2026-10-06
Distributed Systems
Lesson
Hinted Handoff: Keeping Writes Available During Node Failures
When one of the natural replicas for a key is unreachable, a quorum store does not have to fail the write. The node accepting the write durably records a hint — target replica, key, mutation, version
machine-learning
2026-10-05
Algorithms & Data Structures
Lesson
Stable vs Unstable Sorting: Why Equal Items Keep Their Order
All records sharing one key. Stability has nothing to say about records with different keys — only about who keeps their place inside a group of equals.
machine-learning
2026-10-05
AI/ML
Lesson
Airflow vs. Prefect: Cron Schedules vs Event-Driven Flows
Two answers to the same two questions - who decides the shape, what starts the run
machine-learning
2026-10-04
AI/ML
Lesson
Why Random Search Beats Grid Search at the Same Budget
machine-learning
2026-10-04
Engineering & Growth
Lesson
The Number That Tells You How Fragile Your Team Really Is
The bus factor is a measure, not a principle . It counts the minimum number of people who would need to leave a project — resignation, illness, transfer, promotion — before the project stalls due to k
machine-learning
2026-10-03
System Design
Lesson
The Vine That Replaces The Tree Without Cutting It Down
Build the whole replacement first, then flip one switch. Forces a feature freeze on the old system, a binary cut-over date where both must be perfect, and a rollback that is just "put the old one back
machine-learning
2026-10-03
AI/ML
Lesson
AI Security: Untrusted Text Is Data, Not Instructions
The model can be persuaded. Your controls can't.
machine-learning
2026-10-02
LLM & Inference
Lesson
The Difference Between Telling And Showing A Model What You Want
Every prompt is an instruction, but a model's interpretation of that instruction can drift in ways the words alone can't control. Telling (zero-shot) hands it only the instruction — the model guesses
machine-learning
2026-10-02
Security & Identity
Lesson
GRC Engineering: Turning Risk, Policy, and Compliance into Decisions
Four roles. One chain. One test that proves it.
machine-learning
2026-10-01
Career & Learning Craft
Lesson
Transfer: Why a Skill You Own Fails in a New Context
machine-learning
2026-10-01
Distributed Systems
Lesson
How Two Machines Agree On A Number Without Talking
machine-learning
2026-09-30
Distributed Systems
Lesson
The Node That Keeps Answering And Lies
The assumption about how a component may misbehave. Crash or lie — the choice decides which protocol family you are even allowed to run.
machine-learning
2026-09-30
AI/ML
Lesson
RLHF vs DPO: How Preference Data Replaces the Reward Model
machine-learning
2026-09-29
AI Safety & Robustness
Lesson
Proving One Person’s Answer Could Not Have Changed the Result
Every older privacy idea tried to scrub the data \u2014 strip the names, bucket the ages, k-anonymise the rows. All of them failed the same way: an attacker who joins your "anonymous" release against
machine-learning
2026-09-29
AI/ML
Lesson
Goodhart's Law in ML: When a Metric Becomes a Target
machine-learning
2026-09-28
Databases & Storage
Lesson
Two Transactions Each Holding What the Other Needs
One node per transaction, an arrow A→B when A waits on a lock B holds. A closed cycle = a deadlock, and nothing else qualifies.
machine-learning
2026-09-28
AI/ML
Lesson
Draft-Verify Inference: One Parallel Pass, Five Candidate Tokens
Sticky note - the whole lesson: One weight stream per step, k tokens from idle compute.
machine-learning
2026-09-27
Incident Response & Operations
Lesson
The One Field That Turns Ten Log Files Into One Story
Correlation IDs, spans & the traceparent header — the mechanism that makes a cross-service failure reconstructable
machine-learning
2026-09-27
AI/ML
Lesson
Reflex to Learning: The Classic Taxonomy of AI Agent Architectures
Five rungs, one force - each level fixes the one below.
machine-learning
2026-09-26
Retrieval & Vector Search
Lesson
Why a Chunk Forgets the Document It Came From
Standard RAG embeds each chunk on its own . The chunk keeps its words and loses everything the document knew about it — the company, the year, the product. Its embedding is therefore a vector for a ge
machine-learning
2026-09-26
AI/ML
Lesson
What is normalization?
machine-learning
2026-09-25
Networking & Traffic Management
Lesson
The First Visitor Pays So the Next Thousand Do Not
A content delivery network (CDN) is a fleet of caches “edges” parked in many cities, each sitting between your visitors and your origin server. When a request arrives at an edge that has never seen th
machine-learning
2026-09-25
Engineering & Growth
Lesson
Rate Limiters
429 Too Many Requests: how a token bucket decides who waits and who passes. Show: burst allowance, refill rate, and what happens when a client hammers the API. Include the X-RateLimit-Remaining header
machine-learning
2026-09-25
AI/ML
Lesson
Non-Maximum Suppression: Deduplicating Bounding Boxes with IoU
One car, three boxes. NMS keeps the best. Panel one - The over-detection problem: - One object, many overlapping boxes - Each box carries a confidence score - Keep the top-confidence box - Discard sig
machine-learning
2026-09-24
Algorithms & Data Structures
Lesson
The Cheapest Way to Connect Everything Without Going in Circles
For any cut of the vertices, the lightest crossing edge is in some MST. This is the entire proof that the cheapest available edge can never be a mistake.
machine-learning
2026-09-24
System Design
Lesson
One Writer, Fifty Readers: How Far a Single-Primary Database Actually Goes
Subtitle: One writer. Fifty readers. All load on one box.
machine-learning
2026-09-24
AI/ML
Lesson
Context Engineering: The Orchestration Layer That Decides What the Model Sees
Prompting tunes words; context engineering tunes the model's evidence.
machine-learning
2026-09-23
Networking & Traffic Management
Lesson
The Balancer That Reads Your Request and the One That Refuses To
Reads only the TCP 5-tuple. Chooses a backend when the connection opens; has no idea where one request ends and the next begins.
machine-learning
2026-09-23
AI/ML
Lesson
Context Engineering: The Orchestration Layer That Decides What the Model Sees
Prompting tunes words; context engineering tunes the model's evidence.
machine-learning
2026-09-23
System Design
Lesson
Anti-Entropy: The Background Sweep That Compares Whole Replicas, Not Single Reads
Subtitle: Swap the roots first - pay only for the difference. Panel one - The gap only it fills: - Read repair: read path only - Hinted handoff: hints expire - Neither touches cold keys - Only repair
machine-learning
2026-09-22
AI/ML
Lesson
Context Engineering: The Orchestration Layer That Decides What the Model Sees
Prompting tunes words; context engineering tunes the model's evidence.
machine-learning
2026-09-21
AI/ML
Lesson
Context Engineering: The Orchestration Layer That Decides What the Model Sees
Prompting tunes words; context engineering tunes the model's evidence.
machine-learning
2026-09-20
Databases & Storage
Lesson
The Level Names Tell You What Cannot Happen, Not How It Works
READ COMMITTED is not a description of an implementation. It is a description of a thing that cannot happen to you . The four isolation levels are not four algorithms — they are four promises about wh
machine-learning
2026-09-20
Career & Learning Craft
Lesson
Cybersecurity Certifications and Portfolio: Evidence Beyond Badges
Subtitle: Projects prove value. Badges only promise.
machine-learning
2026-09-19
AI Safety & Robustness
Lesson
The Pixel Change You Cannot See That Breaks the Model
A trained classifier draws decision boundaries in a space with far more dimensions than the data actually fills — and in high dimensions, almost every point sits close to a boundary , even when it loo
machine-learning
2026-09-19
Security & Identity
Lesson
Application Security: Integrating SAST, DAST, Code Review, and DevSecOps
Taped rule strip: Tool count is not coverage.
machine-learning
2026-09-18
Security & Identity
Lesson
Cybersecurity Foundations: CIA Triad, Threat Models, and Defense in Depth
machine-learning
2026-09-17
AI/ML
Lesson
CAI: Constitution-Based Critique and Revision
Subtitle: Written rules a model applies to itself. Left column - The supervised loop: - Draft an answer - Name the principle it breaks - Revise, keep it useful - Critique the revision again - Accept o
machine-learning
2026-09-16
System Design
Lesson
Streaming Fetch: Reading response.body Before the Download Completes
Left column - The habit that blocks: - fetch() settles once headers arrive - response.json() waits for every byte - .text() too - one full string first - So the UI sees nothing until the end - Streami
machine-learning
2026-09-15
AI Platform & MLOps
Lesson
Turning an Hour of Speech Into Text You Can Search
Managed ASR (Automatic Speech Recognition) — AWS Transcribe is the reference example — takes audio you place in object storage (S3) and returns JSON, not a paragraph . Every single token in that JSON
machine-learning
2026-09-15
System Design
Lesson
Cache Invalidation at Scale: Versioned Keys, Purges, and Bounded Staleness
machine-learning
2026-09-14
AI Algorithms Curriculum
Lesson
One Word With One Vector Cannot Mean Two Things
word2vec and GloVe assign each word ONE vector, learned from which words co-occur across a corpus. That single choice is the whole story.
machine-learning
2026-09-14
System Design
Lesson
FOIT and font-display: swap: Faster Text with a Fallback Flash
Subtitle: Read now; swap later without losing layout.
machine-learning
2026-09-13
AI/ML
Lesson
Random Seeds in LLM Sampling: Reproducibility Without a Guarantee
Subtitle: A seed can replay a draw. It cannot freeze the whole stack.
machine-learning
2026-09-13
Web & Browser APIs
Lesson
Putting a Hologram on Your Table Through a Web Page
WebXR is the browser API that replaced WebVR, and its core job is exactly one thing: give your code a POSE every single frame — where the headset or phone is, expressed in a coordinate system the real
machine-learning
2026-09-12
Algorithms & Data Structures
Lesson
Taking the Best Bite Now and Never Looking Back
A greedy algorithm makes the locally optimal choice at each step and never revises it . There is no backtracking, no exploring of alternatives, no memo table. That is the whole design — and it is why
machine-learning
2026-09-12
System Design
Lesson
LRU vs LFU: How Caches Decide What to Evict
Subtitle: Recency and frequency protect different reuse patterns.
machine-learning
2026-09-11
AI Platform & MLOps
Lesson
Where the Model Actually Runs When Someone Hits Your Endpoint
A trained model does nothing until something asks it a question. Serving is the choice of when that question gets answered — the instant it is asked, on a schedule before it is asked, or as events flo
machine-learning
2026-09-11
AI/ML
Lesson
The Difference Between Telling and Showing a Model What You Want
Subtitle: Instructions name the task. Examples reveal its shape.
machine-learning
2026-09-10
Data Engineering
Lesson
The Pipe That Batches Your Events So the Warehouse Does Not Choke
This is a delivery pipe, not a queue you read from — and that single distinction is the whole point. You write records in . It buffers them, and on its own schedule it writes them out to a destination
machine-learning
2026-09-10
System Design
Lesson
How Two Machines Agree on a Number Without Talking
Subtitle: Snowflake IDs replace agreement with bit ownership.
machine-learning
2026-09-09
System Design
Lesson
The Node That Keeps Answering, and Lies
Subtitle: Silence is easy to spot. A convincing lie is not.
machine-learning
2026-09-09
AI Platform & MLOps
Lesson
A Catalogue So You Know Which Agents Exist Before They Bill You
An agent registry is a control plane for models, agents and their tools: a single inventory recording which agents exist , which model version each one calls, what permissions and data it may touch, a
machine-learning
2026-09-08
Web & Browser APIs
Lesson
Why Scrolling a Small Box Drags the Whole Page With It
Scroll chaining, and the one CSS line that refuses it.
machine-learning
2026-09-08
System Design
Lesson
Why One Cheap Shortcut Backwards Breaks the Whole Map
Subtitle: Smallest-first is a proof, not a habit.
machine-learning
2026-09-07
Security & Identity
Lesson
The Key Your Own Code Is Not Allowed to Read
Subtitle: A key can be usable without being readable.
machine-learning
2026-09-07
AI Algorithms Curriculum
Lesson
Teaching a Model What Good Looks Like When You Cannot Write It Down
You can write a loss function for "predict the next token" . You cannot write one for "helpful and not rude" . RLHF's move is to stop trying: instead of specifying the objective, you learn it from hum
machine-learning
2026-09-06
AI/ML
Lesson
You Pay for Training Once and for Answering Forever
1. SAME MODEL, DIFFERENT WORK — Training: forward pass, loss, backprop, gradients, weight update; stores batch activations and repeats across epochs. Inference: forward pass only; no gradients, histor
machine-learning
2026-09-06
Web & Browser APIs
Lesson
Blurring What Is Behind an Element Instead of the Element Itself
filter: blur() blurs the element you put it on — its own text, its own borders, its own children. That is almost never what you wanted, because you wanted the panel readable .
machine-learning
2026-09-05
System Design
Lesson
The Meeting Where Nobody Should Be Named
Subtitle: A post-mortem is a shape, not a mood.
machine-learning
2026-09-05
Cloud & Infrastructure
Lesson
The Backend You Did Not Write and Still Have to Run
Amplify is not one service, and there is no Amplify runtime hiding underneath . It is a code generator plus a deployment pipeline sitting over ordinary AWS parts.
machine-learning
2026-09-04
System Design
Lesson
The Structure That Only Ever Knows Its Smallest Item
Subtitle: One weak rule. One cheap answer.
machine-learning
2026-09-04
Engineering & Growth
Lesson
Crashing Early on Purpose So You Never Guess Later
Two halves, pulling in opposite directions. Fail fast says: stop dead the instant an assumption is violated. Fallback grace says: keep serving something rather than propagating a failure to the user.
machine-learning
2026-09-03
Web & Browser APIs
Lesson
The Collapsible Panel That Needs No JavaScript at All
The <details> element and its <summary> child form a native HTML disclosure widget. The browser handles the toggle, exposes an open boolean attribute you can read, set, and style against, fires a togg
machine-learning
2026-09-02
System Design
Lesson
The Function Call That Can Run Twice
Subtitle: It reads like local code. It is a network.
machine-learning
2026-09-02
Career & Learning Craft
Lesson
How to Lose an Argument and Still Ship on Monday
Disagree and commit is a decision rule with two halves, and it only works when both are present. First: your disagreement gets recorded before the decision — in the open, with the reasoning, and genui
machine-learning
2026-09-01
System Design
Lesson
What One More Nine Actually Costs You in Minutes
- Compute your nines from raw request logs, per region. Compare that with the dashboard headline.
machine-learning
2026-09-01
Career & Learning Craft
Lesson
The Document That Makes Disagreement Cheap
State the problem and its evidence first — quantities, not adjectives. Then the constraints you must respect. A proposal that opens with its solution is asking for agreement, not review.
machine-learning
2026-08-31
LLM & Inference
Lesson
Why the Bill Counts Two Kinds of Token
A completion token is a token the model generated . A prompt token is one it read . Providers bill them at different rates — typically 2× to 4× more for completion — and the asymmetry is not arbitrary
machine-learning
2026-08-31
System Design
Lesson
When Two Writes Must Stay in Order and the Rest May Not
Subtitle: Only linked writes are fixed in order. Left column - Causally related (order is fixed): - You read a value, then you write - Your write depends on that read - Every replica shows that order
machine-learning
2026-08-30
LLM & Inference
Lesson
The Wait Before the First Word Appears
Time to first token (TTFT) is the latency from sending a request to receiving the first streamed token. It is dominated by prefill : the single pass in which the model processes every prompt token in
machine-learning
2026-08-30
System Design
Lesson
A System That Is Always Up and Frequently Wrong
Subtitle: Uptime is not the same as being right. Left column - Availability - did it answer?: - Uptime: something came back - Counts responses, not truth - A stale cache scores high here - Redundancy
machine-learning
2026-08-29
System Design
Lesson
Cutting a Shape Out of a Rectangle the Browser Insists On
Subtitle: Clipping changes the picture, not the box. Left column - Layout: still a full rectangle: - Every element laid out is a box - It still occupies that rectangle - It pushes neighbours the same
machine-learning
2026-08-29
LLM & Inference
Lesson
Asking a Weaker Question First to Get a Better Answer
You have a hard, detail-heavy question. The instinct is to ask it directly. Step-back prompting says: don't . First ask a deliberately weaker, more general question — what principle governs this kind
machine-learning
2026-08-28
Cloud & Infrastructure
Lesson
The Copy That Does Not Happen Until You Write
Several processes share one instance of the data instead of each getting a private copy. The duplicate is created at the exact moment a process tries to modify it — not a moment sooner.
machine-learning
2026-08-28
System Design
Lesson
Ordering Events Without Trusting Any Clock
Subtitle: Throw away the clock. Keep a counter. Left column - Wall clock timestamps: - Two machines never agree exactly - Clocks drift apart every hour - A sync can jump time backwards - So a smaller
machine-learning
2026-08-27
System Design
Lesson
Write-Ahead Logging
Before a storage engine touches a data page, it appends a description of the change to a sequential log and forces that log to stable storage. The data pages themselves are updated later, lazily, in a
machine-learning
2026-08-27
System Design
Lesson
Choosing a Data Centre Before Choosing a Server
Subtitle: Two layers, two jobs, one request.
machine-learning
2026-08-27
System Design
Lesson
Bloom Filters
INSERT: for the key "apple", the three hash functions produce positions 2, 7 and 19. Set bits at positions {2, 7, 19} to 1. EXACTLY three cells of the 24 are filled -- cell 2, cell 7 and cell 19 -- an
machine-learning
2026-08-26
System Design
Lesson
The Outline That Appears for the Keyboard and Hides for the Mouse
Subtitle: Same ring - only the keyboard sees it.
machine-learning
2026-08-25
AI/ML
Lesson
Searching With Small Pieces and Answering With Big Ones
Subtitle: Index one sentence, return its neighbours.
machine-learning
2026-08-25
System Design
Lesson
Bloom Filters
A Bloom filter trades certainty for space . Instead of storing the items themselves, it stores a fingerprint of them in a fixed bit array — then answers one question only: is this item in the set?
machine-learning
2026-08-24
System Design
Lesson
The Copy That Does Not Happen Until You Write
Subtitle: Share one page until someone writes.
machine-learning
2026-08-24
System Design
Lesson
Circuit Breaker Pattern
When a downstream dependency gets slow or starts erroring, the naive caller keeps sending requests and keeps waiting on timeouts. Each waiting request holds a thread, a connection, a chunk of memory.
machine-learning
2026-08-23
System Design
Lesson
Why Shifting Right Is Not the Same as Dividing
Subtitle: They agree on positives. Not on signs.
machine-learning
2026-08-23
System Design
Lesson
CQRS — Command Query Responsibility Segregation
CQRS splits the single data model that normally serves both writes and reads into two independent models : a command model that changes state and returns nothing, and a query model that returns data a
machine-learning
2026-08-22
System Design
Lesson
One Endpoint That Returns Exactly the Fields You Asked For
Subtitle: One schema, many backends, one shape.
machine-learning
2026-08-22
AI Algorithms Curriculum
Lesson
Event Sourcing
A normal database stores the current answer and throws away the question. Event Sourcing inverts that: the only source of truth is an ordered, append-only log of facts that already happened — and curr
machine-learning
2026-08-21
System Design
Lesson
Why a Log Is a Better Queue Than a Queue
Subtitle: One takes it away. One remembers. Left column - Queue (delete on read): - Message consumed, then deleted - Second reader gets nothing - No rewind, no replay - Broker tracks work left to do -
machine-learning
2026-08-21
AI Algorithms Curriculum
Lesson
CAP Theorem
In a distributed data store, when the network between nodes drops messages — a partition — you cannot simultaneously keep every read consistent (all nodes agree on the latest write) and keep the syste
machine-learning
2026-08-20
AI/ML
Lesson
Learning From Data With No Labels
Subtitle: The labels come from the data.
machine-learning
2026-08-20
AI Algorithms Curriculum
Lesson
Consistent Hashing
The naive way to shard data is server = hash(key) mod N . It works perfectly until N changes. Go from 4 servers to 5 and almost every key's answer changes — roughly (N-1)/N of your data , about 80%, m
machine-learning
2026-08-19
Security & Identity
Lesson
Why One Ampersand Instead of Two Is a Real Bug
Use it when - Bitwise AND: permission masks, device flags
machine-learning
2026-08-19
AI Algorithms Curriculum
Lesson
Beam Search
Beam search is a bounded breadth-first walk over the tree of possible output sequences.
machine-learning
2026-08-18
System Design
Lesson
Why Your App Shows Stale Data Even Though the Write Succeeded
Subtitle: The write was fine. The read never promised to see it.
machine-learning
2026-08-17
AI Algorithms Curriculum
Lesson
Seq2Seq
Classifiers map an input to one label . Seq2Seq maps an input sequence of length T to an output sequence of length T′ — and T′ need not equal T. An encoder RNN/LSTM reads the whole input and squeezes
machine-learning
2026-08-17
Security & Identity
Lesson
Hashing vs Encryption vs Encoding
Use it when - ENCRYPTION: secrets in transit or at rest Use it when - ENCODING: binary in JSON, URLs, email
machine-learning
2026-08-16
AI Algorithms Curriculum
Lesson
Word2Vec
Word2Vec (Mikolov et al., 2013) takes that linguistic intuition and makes it a prediction task . Train a deliberately shallow network to predict a word from its context (or its context from the word),
machine-learning
2026-08-16
System Design
Lesson
Scaling: Vertical vs Horizontal vs Auto
Subtitle: Add power, add boxes, or add rules. Left column - VERTICAL (scale up): Bigger machine: more CPU and RAM No code change needed One box, one hard ceiling Resize usually means downtime Cost cli
machine-learning
2026-08-15
Career & Learning Craft
Lesson
Dual Coding: Words + Visuals
Subtitle: Your brain has two channels. Use both. Left column - Words only: One channel carries everything Text and talk share one path Feels fast, fills up fast Load hits a limit, detail drops Recall
machine-learning
2026-08-14
AI Algorithms Curriculum
Lesson
Monte Carlo Tree Search
Every node stores just two numbers: n (times visited) and w (total reward backed up through it). No board evaluation, no learned weights — two counters.
machine-learning
2026-08-14
Security & Identity
Lesson
CORS vs CSRF
Subtitle: One blocks reads. One forges writes. Left column - CORS: Browser rule for cross-site reads Server sends Allow-Origin header Blocks the page, not the request Guards your data from other sites
machine-learning
2026-08-13
AI Algorithms Curriculum
Lesson
Proximal Policy Optimization
Improve the policy as much as you can — but never trust one update too far.
Vanilla policy gradient has one fatal habit: a single large step can destroy a working policy, and there is no way back — the data that would correct it was collected by the policy you just wrecked. P
machine-learning
2026-08-13
System Design
Lesson
Monolith vs SOA vs Microservices vs Event-Driven
Subtitle: One axis explains all four - coupling.
machine-learning
2026-08-12
Career & Learning Craft
Lesson
Worked Examples vs Solving Cold
Subtitle: When to study the solution and when to struggle first Left column - WORKED EXAMPLES: See every step already solved Low mental load, room to see pattern Best when the topic is brand new Gives
machine-learning
2026-08-12
AI Algorithms Curriculum
Lesson
Actor-Critic Methods
Pure policy gradient (REINFORCE) is unbiased but wildly noisy — it waits for a whole episode's return before it learns anything, and that return swings hard. Pure value methods (Q-learning) are stable
machine-learning
2026-08-11
Security & Identity
Lesson
Passkeys vs Passwords: WebAuthn Basics
Subtitle: A shared secret vs a private key that never leaves your device
machine-learning
2026-08-11
AI Algorithms Curriculum
Lesson
CatBoost
CatBoost is gradient boosting on decision trees, engineered around one observation: the standard ways of handling categorical features leak the target into the features . Naive mean-target encoding co
machine-learning
2026-08-10
System Design
Lesson
The Consistency Ladder: Eventual to Strong
Subtitle: Five levels of read guarantee, weakest to strongest.
machine-learning
2026-08-10
AI Algorithms Curriculum
Lesson
LightGBM
LightGBM is gradient boosting where every expensive step has been replaced by a cheaper approximation that barely costs accuracy. Like all boosting, it fits trees sequentially, each one predicting the
machine-learning
2026-08-09
AI Algorithms Curriculum
Lesson
Spectral Clustering
k-Means draws straight lines . That is its whole geometry — every cluster is the set of points nearest one centroid, so every boundary is a flat hyperplane. Give it two interlocking crescent moons and
machine-learning
2026-08-09
Career & Learning Craft
Lesson
Elaborative Interrogation: Ask Why, Not What
Subtitle: Turn facts into reasons you can rebuild
machine-learning
2026-08-08
AI Algorithms Curriculum
Lesson
UMAP
UMAP assumes your high-dimensional data actually lives on a much lower-dimensional manifold — a curved sheet folded up inside a big space. It never trusts long distances. Instead it asks each point, "
machine-learning
2026-08-08
Security & Identity
Lesson
TOTP: How Authenticator Codes Work
Subtitle: A shared secret plus a shared clock, no network needed
machine-learning
2026-08-07
AI Algorithms Curriculum
Lesson
Expectation Maximization
Maximum likelihood is easy when you can see everything. EM is what you reach for when some of the data is missing by design — the cluster label, the regime, the hidden state was never recorded.
machine-learning
2026-08-07
System Design
Lesson
Back-of-the-Envelope Estimation
Subtitle: Get to a defensible number in five minutes
machine-learning
2026-08-06
Career & Learning Craft
Lesson
The Generation Effect: Guess Before You're Told
Subtitle: Struggle for the answer before you read it
machine-learning
2026-08-06
AI Algorithms Curriculum
Lesson
Hidden Markov Models
A Hidden Markov Model describes a system that moves through a sequence of hidden states you never observe directly. What you do observe are emissions — noisy signals produced by whichever state the sy
machine-learning
2026-08-05
Security & Identity
Lesson
Refresh Token Rotation
Subtitle: Short-lived access, single-use refresh, revoke the family on reuse
machine-learning
2026-08-05
AI Algorithms Curriculum
Lesson
Gaussian Processes
Encodes “how similar are two inputs?” and thereby every assumption about the function. RBF: σ²exp(−(x−x′)²/2ℓ²) . The lengthscale ℓ is the wiggle rate; small ℓ means twitchy, large ℓ means smooth. Ker
machine-learning
2026-08-04
System Design
Lesson
Observability: Logs vs Metrics vs Traces
Subtitle: Three signals, three questions, one system
machine-learning
2026-08-04
AI Algorithms Curriculum
Lesson
Bayesian Optimization
Grid search asks every question. Random search asks arbitrary questions. Bayesian Optimization asks the most informative next question — and typically finds a good optimum in tens of evaluations where
machine-learning
2026-08-03
Career & Learning Craft
Lesson
Feedback Loops: Tight vs Loose
Subtitle: How fast you learn is how fast you find out
machine-learning
2026-08-03
AI Algorithms Curriculum
Lesson
Federated Learning
machine-learning
2026-08-02
Security & Identity
Lesson
OAuth 2.0 Authorization Code Flow with PKCE
Subtitle: How a public app proves it started the login
machine-learning
2026-08-02
AI Algorithms Curriculum
Lesson
Contrastive Learning
Random crop, colour jitter, grayscale, blur, flip. This is the supervision signal — weak augmentation makes the task trivial, so the model learns nothing but colour histograms.
machine-learning
2026-08-01
AI Algorithms Curriculum
Lesson
Graph Neural Networks
A CNN assumes your data sits on a grid. An RNN assumes it sits on a line. A GNN drops both assumptions: it learns on arbitrary graphs , where each node has features and the edges say who is related to
machine-learning
2026-08-01
System Design
Lesson
Failure Handling: The Four Moves
Subtitle: Four moves that stop one slow service from sinking everything
machine-learning
2026-07-31
AI Algorithms Curriculum
Lesson
Diffusion Models
A diffusion model learns generation by learning destruction in reverse . Take a real image and add a little Gaussian noise, again and again, over T steps — eventually it is pure static. That forward p
machine-learning
2026-07-31
Career & Learning Craft
Lesson
Deep Work vs Shallow Work
Subtitle: Protect attention first, batch the rest
machine-learning
2026-07-30
Security & Identity
Lesson
Session Cookies vs JWT
Subtitle: Stateful memory vs stateless proof
machine-learning
2026-07-30
System Design
Lesson
PACELC: Latency vs Consistency
Subtitle: Latency and consistency are the everyday trade
machine-learning
2026-07-30
AI Algorithms Curriculum
Lesson
GPT Architecture
GPT throws away half of the original Transformer. No encoder, no cross-attention — just a tall stack of identical decoder blocks reading left to right. Every position may attend to itself and everythi
machine-learning
2026-07-29
Career & Learning Craft
Lesson
Retrieval Practice vs Re-reading
Subtitle: Pull it out of your head, do not push it in again
machine-learning
2026-07-29
AI Algorithms Curriculum
Lesson
BERT
BERT takes the Transformer and throws away the decoder. What remains is a deep encoder stack that reads an entire sentence in both directions at once — so the vector for a word is shaped by everything
machine-learning
2026-07-28
System Design
Lesson
Raft Leader Election
Subtitle: How a cluster agrees on exactly one leader
machine-learning
2026-07-28
AI Algorithms Curriculum
Lesson
Attention Mechanisms
Let every position look back at every other position — and weight what actually matters.
machine-learning
2026-07-27
Security & Identity
Lesson
Password Hashing: bcrypt vs Argon2id
Subtitle: Store passwords so stolen data stays useless
machine-learning
2026-07-27
Engineering & Growth
Lesson
Handle watermark verification
machine-learning
2026-07-27
Engineering & Growth
Lesson
Gradient descent intuition
machine-learning
2026-07-27
Engineering & Growth
Lesson
Generative Adversarial Networks (GANs)
Educational hand-drawn sketchnote explaining Generative Adversarial Networks as a two-player adversarial game. Center: a large loop diagram — random noise vector z (small dice/scribble cloud) feeding
machine-learning
2026-07-27
AI Algorithms Curriculum
Lesson
Generative Adversarial Networks
Most generative models ask you to write down a likelihood and maximise it. GANs refuse. Instead they train a Generator G that maps random noise z to samples, and a Discriminator D that scores samples
machine-learning
2026-07-26
AI Algorithms Curriculum
Lesson
Variational Autoencoders (VAE)
Learn a latent distribution instead of a single point, so the space between examples becomes generative.
machine-learninggenerativedeep-learningvae
2026-07-26
System Design
Lesson
Load Balancer vs API Gateway
Subtitle: Traffic distribution vs traffic control
machine-learning
2026-07-26
Career & Learning Craft
Lesson
Interleaving vs Blocked Practice
Subtitle: Mix your reps to build real judgment
machine-learning
2026-07-25
AI Algorithms Curriculum
Lesson
🧬 Genetic Algorithms
An encoded candidate solution — often a bitstring or vector of ‘genes’ representing the parameters being optimized.
machine-learning
2026-07-25
Security & Identity
Lesson
Mutual TLS (mTLS): Two-Way Trust
1) What mTLS Is: Standard TLS proves only the SERVER's identity to the client (one-way). mTLS adds the reverse: the CLIENT also presents a certificate, so BOTH sides prove who they are before any data
machine-learning
2026-07-24
AI Algorithms Curriculum
Lesson
Markov Decision Process (MDP)
machine-learning
2026-07-23
Career & Learning Craft
Lesson
The Feynman Technique
THE CORE IDEA: You do not understand something until you can explain it simply. The Feynman Technique turns passive review into active generation \u2014 you learn by teaching an imaginary novice, and
machine-learning
2026-07-23
AI Algorithms Curriculum
Lesson
Isolation Forest
machine-learning
2026-07-22
System Design
Lesson
Database Indexing
WHY INDEXES EXIST: A table scan reads every row (O(n)); an index is a sorted side-structure that turns lookups into O(log n). Trade-off: faster reads, slower writes, more storage — the librarian's car
machine-learning
2026-07-22
AI Algorithms Curriculum
Lesson
Autoencoders
Maps input x to a compressed code: f(x) = z. Layers shrink toward the bottleneck.
machine-learning
2026-07-21
AI Algorithms Curriculum
Lesson
k-Means++
Smart seeding that makes clustering start in the right place.
machine-learning
2026-07-21
System Design
Lesson
Rate Limiting Algorithms
Rate Limiting Algorithms — engineering visual from the Sensei rich-content series. Tap the infographic to download the full-resolution version.
machine-learning
2026-07-20
AI Algorithms Curriculum
Lesson
Transformer
The Transformer replaces recurrence with self-attention : every token looks at every other token in one parallel pass and decides how much to weigh each. No sequential loop means the whole sequence is
machine-learning
2026-07-19
AI Algorithms Curriculum
Lesson
Long Short-Term Memory
Plain RNNs multiply gradients through time, so error signal either vanishes or explodes over long sequences. The LSTM adds a protected cell state that information flows through with mostly additive up
machine-learning
2026-07-18
AI Algorithms Curriculum
Lesson
Recurrent Neural Network
An RNN is a neural network with a loop . At each step it reads one element of a sequence and blends it with a hidden state — a running memory of everything seen so far. That memory is what lets it mod
machine-learning
2026-07-17
Security & Identity
Lesson
JWT Security
JWT Security — engineering visual from the Sensei rich-content series. Tap the infographic to download the full-resolution version.
machine-learning
2026-07-17
AI Algorithms Curriculum
Lesson
Convolutional Neural Network
machine-learning
2026-07-16
AI Algorithms Curriculum
Lesson
Artificial Neural Network
The learnable parameters. Weights scale each input's influence; the bias shifts the activation threshold. Training = finding good values for these.
machine-learning
2026-07-15
AI Algorithms Curriculum
Lesson
Actor-Critic
Maps states to a distribution over actions. Updated in the direction the Critic says is advantageous.
machine-learning
2026-07-14
Career & Learning Craft
Lesson
Spaced Repetition
Spaced Repetition — engineering visual from the Sensei rich-content series. Tap the infographic to download the full-resolution version.
machine-learning
2026-07-14
AI Algorithms Curriculum
Lesson
Policy Gradient — REINFORCE
A network mapping state → action probabilities. Differentiable, so we can push gradients through it.
machine-learning
2026-07-13
AI Algorithms Curriculum
Lesson
Deep Q-Network (DQN)
machine-learning
2026-07-12
AI Algorithms Curriculum
Lesson
SARSA — On-Policy TD Control
(s, a, r, s', a') — the update needs the next action a' chosen by the current policy, not the max over actions.
machine-learning
2026-07-11
AI Algorithms Curriculum
Lesson
Q-Learning
The agent observes a state s , picks an action a , and gets reward r plus a next state s' . The (s,a,r,s') tuple is the unit of learning.
machine-learning
2026-07-10
AI Algorithms Curriculum
Lesson
t-SNE
High-D closeness modeled as a Gaussian probability — how likely point i picks j as a neighbor.
machine-learning
2026-07-10
System Design
Lesson
Caching Strategies
Caching Strategies — engineering visual from the Sensei rich-content series. Tap the infographic to download the full-resolution version.
machine-learning
2026-07-09
AI Algorithms Curriculum
Lesson
🎯 Principal Component Analysis
machine-learning
2026-07-08
AI Algorithms Curriculum
Lesson
DBSCAN
The neighborhood radius. How close two points must be to count as neighbors.
machine-learning
2026-07-07
AI Algorithms Curriculum
Lesson
Hierarchical Clustering
Hierarchical clustering builds a hierarchy of groups rather than a flat partition. The agglomerative (bottom-up) version starts with every point as its own cluster and repeatedly merges the two closes
machine-learning
2026-07-07
Career & Learning Craft
Lesson
Deliberate Practice
Deliberate Practice — engineering visual from the Sensei rich-content series. Tap the infographic to download the full-resolution version.
machine-learning
2026-07-06
AI Algorithms Curriculum
Lesson
k-Means Clustering
k-Means partitions data into k groups by minimizing the total within-cluster variance. Each point joins the cluster whose centroid (mean) is nearest, and centroids are re-computed until they stop movi
machine-learning
2026-07-05
AI Algorithms Curriculum
Lesson
XGBoost
machine-learning
2026-07-03
Security & Identity
Lesson
OAuth vs OIDC
OAuth vs OIDC — engineering visual from the Sensei rich-content series. Tap the infographic to download the full-resolution version.
machine-learning
2026-07-03
AI Algorithms Curriculum
Lesson
AdaBoost
Simple models (decision stumps) that only need to be slightly better than random guessing.
machine-learning
2026-07-02
AI Algorithms Curriculum
Lesson
Gradient Boosting
Build an ensemble sequentially, each weak tree correcting the residual errors of the trees before it.
machine-learningensembleboostinggradient-boosting
2026-07-01
AI Algorithms Curriculum
Lesson
Naive Bayes
Use Bayes' theorem plus a simplifying independence assumption to build a fast probabilistic classifier.
machine-learningclassificationbayesprobability
2026-06-30
AI Algorithms Curriculum
Lesson
k-Nearest Neighbors (k-NN)
Classify a point by looking at the labels of its nearest examples under a chosen distance metric.
machine-learningclassificationdistanceknn
2026-06-30
Algo & Data Structures
Lesson
Introduction to Algorithms
Understand the fundamental concept of algorithmic complexity and Big O notation for evaluating algorithm performance.
algorithmscomplexitybig-ofundamentals
2026-06-30
System Design
Lesson
Database Sharding
Database Sharding — engineering visual from the Sensei rich-content series. Tap the infographic to download the full-resolution version.
machine-learning
2026-06-29
AI Algorithms Curriculum
Lesson
Support Vector Machine (SVM)
Find the maximum-margin boundary that separates classes, with kernels for non-linear structure.
machine-learningclassificationsvmkernel
2026-06-28
AI Algorithms Curriculum
Lesson
Random Forest
Reduce variance by averaging many decorrelated decision trees trained on bootstrap samples.
machine-learningensemblebaggingrandom-forest
2026-06-27
AI Algorithms Curriculum
Lesson
Decision Tree
Make predictions through interpretable if-then splits chosen by information gain or impurity reduction.
machine-learningtreesentropygini
2026-06-26
AI Algorithms Curriculum
Lesson
Logistic Regression
Turn a linear score into a calibrated probability for binary classification.
machine-learningclassificationsigmoidlog-loss
2026-06-26
Security & Identity
Lesson
API Authentication
API Authentication — engineering visual from the Sensei rich-content series. Tap the infographic to download the full-resolution version.
machine-learning
2026-06-25
AI Algorithms Curriculum
Lesson
Linear Regression
Predict continuous values by fitting the best linear relationship between features and a target.
machine-learningregressionleast-squaresgradient-descent
2026-06-25
Resource Digest
Resource Digest
10 Free AI Learning Platforms
Map high-signal free AI education platforms to the Sensei algorithm curriculum.
resourcesai-learningcurriculumdigest
2026-06-17
System Design
Lesson
Idempotency
Idempotency — engineering visual from the Sensei rich-content series. Tap the infographic to download the full-resolution version.
machine-learning
