Contents
The course library.
65 short courses, 250 hours in all, sorted by the role they prepare you for. Every lesson is free to read.
AI PM
4 courses- 01
AI Technical Fluency for PMs
LLMs, RAG, fine-tuning, agents — what they are, when to use each, and how to scope them.
LLMs for PMs/Prompting & context/RAG vs fine-tuning vs API
Foundational6 lessons4 h 49 min - 02
AI Product Sense & PRDs
The case round and the PRD — product sense for AI, with acceptance criteria for non-deterministic output.
AI product-sense framework/Designing for trust & failure/Prioritization under uncertainty
Intermediate6 lessons4 h 45 min - 03
Eval, Metrics & Responsible-AI Launch
Define success for non-deterministic products, build the eval, and gate the launch like a senior AI PM.
Defining success for AI products/Product KPIs + model-quality metrics/Offline, online & human evaluation
Intermediate6 lessons4 h 48 min - 04
Product Case Studies (Senior Bar)
One case type per lesson — full worked answers so you can recognize and solve every common product-interview shape.
How to run any product case (senior bar)/Case type: Design a 0→1 product/Case type: Improve a metric
Intermediate8 lessons7 h 28 min
Data scientist
4 courses- 05
GenAI Evaluation for Data Scientists
LLM-as-judge, golden sets, faithfulness, and the traps — measure non-deterministic systems you can trust.
Why GenAI eval is hard/Golden datasets & rubrics/LLM-as-judge & its traps
Intermediate6 lessons4 h 44 min - 06
ML Fundamentals for Interviews
Bias-variance, regularization, calibration, and metrics — the modeling round, done rigorously.
Bias-variance & generalization/Loss functions & optimization/Regularization & calibration
Intermediate6 lessons4 h 45 min - 07
Experiment Design & A/B Testing
Hypotheses, power, the peeking problem, and guardrails — experiments that actually hold up.
Hypotheses & metrics/Significance & power/A/B test design & the peeking problem
Intermediate6 lessons4 h 47 min - 08
SQL & Data Modeling Interviews
Windows, joins, and schemas interviewers actually ask - plus production modeling instincts for analytics and product backends.
SQL mental model: grain, order, nulls/Joins & nulls under interview pressure/Aggregations & metric definitions
Intermediate8 lessons6 h 37 min
GTM engineer
4 courses- 09
Customer-Facing AI Prototypes
Build convincing, grounded AI demos for sales and support — fast.
Building a convincing LLM demo, fast/Grounding the demo in the customer’s own documents/Automating sales and support workflows with agents
Intermediate5 lessons4 h - 10
AI Solution Design, Evals & Demo Selling
Architecture deep-dive + the discovery and demo rounds — design and sell enterprise AI.
Enterprise AI system design/Evals for customer-facing AI/Observability & feedback loops
Intermediate6 lessons4 h 46 min - 11
GTM Engineering Stack & Automation
APIs, CRM data, enrichment waterfalls, and automation — the GTM-engineering stack, built right.
RevOps & the GTM funnel, end to end/CRM data modeling & hygiene/APIs, webhooks & auth
Intermediate6 lessons4 h 46 min - 12
GTM Engineering: Outbound Systems
ICP scoring, enrichment waterfalls, sequences, deliverability, reply ops, and attribution — outbound as a production system, not a Zapier pile.
Outbound as a system: funnel, SLAs, unit economics/ICP definition and scoring services that don't rot/Enrichment waterfalls: coverage, cost, and freshness
Intermediate8 lessons6 h 28 min
Software engineer
4 courses- 13
Behavioral & Storytelling for Technical Interviews
STAR that doesn't sound fake — ownership, conflict, failure, and "tell me about a system you built," for builders who never practiced the soft round.
What behavioral rounds actually grade/STAR / PAR that sounds like a senior, not a script/Ownership, agency, and tell me about a project
Intermediate8 lessons6 h 26 min - 14
Coding Interviews for Applied Engineers
Pattern-first DSA for people who ship systems but never grinded LeetCode — narration, production metaphors, and the four-axis hire rubric.
The interview loop and the hire rubric/Two pointers on arrays and lists/Sliding window
Intermediate8 lessons6 h 30 min - 15
Backend Foundations for AI Builders
HTTP, data models, auth, queues, idempotency, and observability — the layer under every LLM feature that interviewers and production both grade.
HTTP, REST, and API design that survives clients/Data modeling, transactions, and when not to normalize/AuthN/Z: sessions, tokens, RBAC, service identity
Intermediate8 lessons6 h 47 min - 16
SWE System Design Interview
The classic 45–60 minute architecture round — primitives, capacity with a why, and five full designs interviewers still grade.
Building blocks every round reuses/The 45–60 min framework + capacity with a why/Multi-region, multi-tenancy, cost, ops
Advanced8 lessons7 h 25 min
Agent engineer
3 courses- 17
Reliable tool contracts for agents
Engineers who can call an LLM API and must let an agent act on a real service.
A tool call is a request, not a result/Retries and the uncertain-write window/Untrusted text cannot grant permission
Intermediate4 lessons2 h 20 min - 18
Agent failure investigation and interview lab
Agent engineers preparing for incident, design, and debugging interviews.
Find the first wrong transition/Build an outcome verifier that cannot be fooled by narration/Defend architecture with a failure budget
Advanced4 lessons2 h 20 min - 19
Stateful agent workflows that survive interruption
Engineers building agents that run across multiple calls, workers, or user approvals.
Persist the task, not just the conversation/Budgets are part of control flow/Parallel work needs independent ownership
Advanced4 lessons2 h 20 min
AI designer
3 courses- 20
AI Design Systems & Patterns
Reusable suggestion, citation, permission, and provenance patterns for AI products.
A pattern language for AI/Naming & terminology/Provenance: authored vs suggested vs executed
Intermediate5 lessons4 h - 21
Prompt & Context as a Design Material
Prototype model-mediated interactions, design AI copy, and define success for AI UX.
How prompts shape UX/Context, retrieval & tools for designers/Designing AI copy & system messages
Intermediate6 lessons4 h 45 min - 22
Designing Human-AI Interaction
Control, transparency, error recovery, trust — design AI features people actually trust.
Mental models for AI/Designing for uncertainty & errors/Trust calibration & transparency
Intermediate6 lessons4 h 52 min
AI engineer
3 courses- 23
Ship Your First LLM Feature
From an API call to a tested, structured, production-grade service.
How LLMs actually behave/Prompting & context engineering/Structured outputs & function schemas
Intermediate6 lessons4 h 48 min - 24
Retrieval-Augmented Generation
Build RAG that actually grounds — and survives production.
Retrieval foundations & the shapes of RAG/Chunking & contextual retrieval/Hybrid retrieval & the multi-stage pipeline
Intermediate7 lessons5 h 30 min - 25
Agents, Evals & LLMOps
Tool-using agents that are bounded, evaluated, observable, and hard to hijack.
Tool & function calling foundations/Agent patterns: plan, act, recover/MCP: standardizing tools & context
Advanced7 lessons5 h 30 min
AI infra platform
3 courses- 26
GPU capacity and inference mechanics
Software engineers moving into model infrastructure.
Build a GPU memory budget/Explain prefill, decode and batching/Choose replication or model parallelism
Intermediate4 lessons1 h 40 min - 27
AI infrastructure interview lab
Candidates who need to defend model infrastructure choices under changing constraints.
Diagnose a training slowdown across ranks/Design a checkpoint that can actually resume/Work a serving incident from observations
Intermediate4 lessons1 h 40 min - 28
Reliable multi-tenant model serving
Engineers who can estimate GPU fit and want to operate shared inference.
Put admission before the expensive work/Make cache reuse measurable and safe/Scale using queue behavior and warm capacity
Intermediate4 lessons1 h 40 min
Backend engineer
3 courses- 29
Asynchronous backend operations
Backend engineers moving long work out of request handlers.
An asynchronous API needs an honest contract/Concurrency is a capacity decision/Retries can turn an outage into overload
Intermediate4 lessons2 h 45 min - 30
Correct writes under failure
Engineers who can build CRUD endpoints and need to defend write correctness.
A successful request is not the invariant/Read the schedule before selecting isolation/A committed order needs a durable next step
Intermediate4 lessons2 h 45 min - 31
Backend debugging and design interviews
Backend candidates who need to explain evidence and trade-offs under interview constraints.
Investigate the payment that appears twice/Review a patch as a future operator/Design one path before scaling the diagram
Advanced4 lessons4 h
Data engineer
3 courses- 32
Event correctness, late data and replay
Engineers who know SQL and need to reason about changing data.
Give each event a business meaning/Draw the exactly-once boundary/Handle late data with explicit window rules
Intermediate4 lessons1 h 40 min - 33
Data engineering interview lab
Candidates who need to solve data cases and defend pipeline decisions.
Reconstruct latest state from a dirty change table/Calculate a cohort metric with a fixed denominator/Design a delayed-event alert pipeline
Intermediate4 lessons1 h 40 min - 34
Trustworthy production data pipelines
Data engineers who need to operate ingestion and analytical outputs.
Join the snapshot to the change log/Evolve schemas without silently changing meaning/Publish data only when its contract passes
Intermediate4 lessons1 h 40 min
DevRel
3 courses- 35
Teach developers through working examples
Engineers moving into Developer Relations and advocates who need stronger technical teaching artifacts.
Start with what the developer must do/Make the demo reproducible on an ordinary laptop/Document the failure a learner will actually see
Intermediate4 lessons4 h - 36
Build a workshop and feedback loop
Developer advocates responsible for workshops, documentation feedback, and adoption evidence.
A workshop is a sequence of learner decisions/Turn workshop friction into a reproducible issue/Measure useful adoption without overclaiming attribution
Intermediate4 lessons4 h - 37
DevRel technical interview studio
Developer Relations candidates preparing for technical demos, writing tasks, and program discussions.
Design a ten-minute technical presentation/Debug the demo without losing the audience/Edit for a successful reader action
Advanced4 lessons4 h
Forward deployed
3 courses- 38
Full-Stack + Enterprise AI Integration
Build the software around model APIs and wire into messy enterprise systems — fast.
Data wrangling under the gun/APIs & integration patterns/The LLM app/agent layer
Intermediate6 lessons4 h 48 min - 39
Discovery, Decomposition & Client Communication
Scope vague problems, run discovery, and manage execs — the client-facing craft.
Discovery: the questions before code/Decomposing vague problems/Defining success & first 90 days
Intermediate6 lessons4 h 50 min - 40
Enterprise RAG & Secure Deployment
Private RAG in a customer VPC — auth, RBAC, observability, guardrails, PII.
Enterprise RAG architecture/Auth, SSO, RBAC & auditability/Private/VPC deployment patterns
Advanced6 lessons4 h 48 min
Founding engineer
3 courses- 41
Ship and operate a first product
Early-team engineers responsible for release, adoption evidence, incidents, and supplier choices.
A release plan includes the previous version/Measure use that changes a customer's work/Operate the incident before writing the postmortem
Intermediate4 lessons4 h - 42
Scope and architecture under uncertainty
Engineers taking product and technical responsibility in a small early team.
A feature request is evidence to investigate/Choose a small architecture that can change/Know the cost of one successful outcome
Intermediate4 lessons4 h - 43
Founding engineer decision interviews
Candidates who must demonstrate product judgment, delivery, and technical ownership in a small team.
Choose what will not fit this week/Leave a decision trail in the work sample/Change the design when the numbers change
Advanced4 lessons4 h
Fullstack engineer
3 courses- 44
Build a reliable async workflow
Full-stack engineers building long-running user workflows across client, API, worker, and storage.
Build the smallest complete export flow/Treat progress messages as hints about durable state/Conflicts and cancellation are user-visible decisions
Intermediate4 lessons4 h - 45
State across browser and server
Engineers who can build a frontend and API but need reliable behavior across both.
One user action crosses several versions of reality/Optimistic interfaces need a reconciliation rule/A server-rendered component is still a data boundary
Intermediate4 lessons4 h - 46
Full-stack interview work samples
Candidates preparing for product engineering work samples and live debugging.
Turn a broad dashboard brief into a finished core/Debug the search that shows the wrong customer/Explain the waterfall with numbers
Advanced4 lessons4 h
LLM engineer
3 courses- 47
ML & LLM System Design Interview
The 60-minute architecture round, worked end to end — framing, data, serving, and the tradeoffs interviewers grade.
The ML system design framework/Data, features & embedding pipelines/Serving, latency & cost at scale
Advanced6 lessons4 h 45 min - 48
Fine-Tuning & Inference Optimization
When to fine-tune, how LoRA/QLoRA and DPO really work, and how to serve a model cheaply at scale.
Fine-tune vs prompt vs RAG — the decision/Datasets for SFT — quality over quantity/LoRA & QLoRA — PEFT and the memory math
Advanced6 lessons4 h 48 min - 49
Transformers & LLM Internals
Attention, tokenization, the KV cache — what actually happens on every forward pass, and why it breaks at scale.
Tokenization deep dive/Attention from scratch/Embeddings & positional encoding
Advanced7 lessons5 h 31 min
ML engineer
3 courses- 50
Training data contracts for ML engineers
ML engineers turning operational events into training and validation datasets.
What did the system know at prediction time?/Split by the unit you want to generalize to/Labels can be missing for a reason
Intermediate4 lessons2 h 20 min - 51
Model release decisions under real constraints
ML engineers responsible for scoring services and model rollout decisions.
A threshold is an operating decision/Ranking quality does not make scores trustworthy probabilities/A release needs an observable escape route
Intermediate4 lessons2 h 20 min - 52
ML interview case lab: debug the decision, not the diagram
ML engineers preparing for quantitative case and debugging interviews.
Use a tiny example to debug the learning loop/Why precision falls when the world changes/Offline improvement can miss the product objective
Advanced4 lessons2 h 20 min
Platform engineer
3 courses- 53
Platform incident and design interviews
Platform candidates who need to work through live incident and design cases.
Separate naming, routing, and application failures/Plan maintenance without confusing it with failure tolerance/Design a safe self-service scaling API
Intermediate4 lessons1 h 40 min - 54
Safe delivery and developer platforms
Platform engineers who build reusable delivery paths for application teams.
Keep a rollout inside its capacity budget/Treat infrastructure state as shared operational data/Build a supported path with an escape route
Intermediate4 lessons1 h 40 min - 55
Kubernetes control and workload contracts
Engineers who can run containers and need to explain cluster behavior.
Read desired state as a control loop/Budget requests, limits and scheduling/Choose probes that do not create an outage
Intermediate4 lessons1 h 40 min
Research engineer
3 courses- 56
Reproducible experiment systems
Research engineers who convert a research idea into a repeatable experiment.
The experiment is larger than the model file/A checkpoint is a continuation contract/A seed is one control, not a reproducibility guarantee
Intermediate4 lessons2 h 20 min - 57
Research engineering interview lab
Research engineers preparing to debug experiments and defend implementation choices.
Reproduction triage: find the first disagreement/Optimize a benchmark without changing the problem/Spend experiment budget on discriminating tests
Advanced4 lessons2 h 20 min - 58
Training performance investigations
Research engineers reducing training time without changing the scientific comparison.
Optimize the critical path you actually measured/Memory accounting before memory tricks/Mixed precision changes arithmetic, so test the update
Advanced4 lessons2 h 20 min
Research scientist
3 courses- 59
Valid empirical claims for research scientists
Research scientists designing and defending empirical ML comparisons.
Name the claim before selecting the experiment/Pair the evidence and respect the sampling unit/Selection changes the meaning of the winning score
Advanced4 lessons2 h 20 min - 60
Research interview lab: defend the evidence
Research scientists preparing for research talks, empirical critique, and proposal interviews.
When the average tells the wrong comparison/Critique a result without substituting your own story/Propose a study that can fail informatively
Advanced4 lessons2 h 20 min - 61
Research study design: isolate the mechanism
Research scientists planning controlled ML studies with limited data and compute.
Ablations need controls that answer the explanation/Equal steps may not mean a fair resource comparison/Keep the evaluation separate from the method's memory
Advanced4 lessons2 h 20 min
Security engineer
3 courses- 62
Secure delivery and useful detection
Security engineers who review build systems and operational controls.
Bind CI identity to the exact deployment authority/Verify what an artifact attestation actually proves/Build a detection rule with a useful evidence trail
Intermediate4 lessons1 h 40 min - 63
Security engineering interview lab
Security candidates who need to review code, investigate evidence and defend risk decisions.
Review an export handler from authority to output/Scope an incident from incomplete audit evidence/Negotiate a temporary security exception
Intermediate4 lessons1 h 40 min - 64
Threat models and authorization boundaries
Software engineers moving into product and infrastructure security.
Turn a feature diagram into a threat model/Make authorization follow the object/Design a server-side fetcher with bounded reach
Intermediate4 lessons1 h 40 min