Available Courses

Build the mental models that distinguish AI application development from ML research.

Understand how language models work: from training to tokens to inference.

Treat context as an engineering budget.

Invoke LLMs with production-shaped API calls, structured output, logging, provider swap, and routing.

Design modular prompts, choose shot strategy and chain-of-thought, and iterate with failure analysis.

Understand embeddings as semantic vectors — token-to-text chain, cosine similarity, chunking, and identifier blind spots.

Connect LLMs to external systems via tools, the ReAct loop, harness validation, and MCP.

Design, debug, and evaluate retrieval-augmented generation pipelines.

Manage the context window as a finite resource: assembly, memory, compaction, and sub-agents.

Understand what the model receives: subword tokens, not words — and why counts vary by model.

Build eval suites, calibrate judges, track safety/cost/latency, instrument traces, and wire CI release gates.

Production RAG upgrades: advanced chunking, query rewriting, Matryoshka dims, web search, RAGAS, and latency tuning.

Use AI coding agents with discipline — steering, PRDs, vertical slices, CI gates, and maintained architectural judgment.

Threat-model LLM apps — prompt injection, RAG poisoning, agent scoping, harness authorization, and adversarial evals in CI.