Track 3
LLM and agent engineering
Context engineering, RAG, tool use, agent orchestration, evals, guardrails, and cost and latency: what AI Engineer and Applied AI posts ask for.
In the live Anthropic and OpenAI boards on 2026-11-19, roughly 40 to 50% of engineering posts mention evals and agents (Sources ยง2.3), and LinkedIn's most common AI Engineer skills are LangChain, RAG and PyTorch (Sources M2). This track covers the engineering, not the research: how to build, measure and operate LLM features.
Frameworks change monthly. Interviewers care that you understand what a framework does underneath, so lessons describe the mechanism and leave the library choice to you.
Building blocks
How LLM features fail, context engineering, RAG end to end, and tool use with MCP.
- 01How LLM features failA working mental model of tokens, context and sampling, organised around the ways production features break.
- 02Prompt and context engineeringTreat the context window as a budget you assemble per request: instructions, examples, retrieved data and state, each earning its tokens.
- 03Retrieval-augmented generation, end to endIngest, chunk, embed, retrieve, rerank, generate with citations, and evaluate retrieval separately from generation.
- 04Tool use, function calling and MCPGive the model typed tools, validate every call, and keep side effects behind permissions you would grant a junior engineer.
Production AI systems
Agent orchestration, evals, guardrails, and cost and latency: what separates a demo from a system.
- 01Agent orchestration: workflows before agentsUse the simplest composition that works: a single call, a fixed workflow, and only then an autonomous agent loop with limits.
- 02Evals: how you know it worksA versioned test set, metrics that match the task, graders you have checked, and a gate on every prompt or model change.
- 03GuardrailsLayered checks on inputs, outputs and actions, chosen from the harms that matter for this product rather than a generic list.
- 04Cost and latencyModel each request in tokens, then pull the levers in order: smaller model, shorter context, caching, batching, streaming.
Sources
- Official Anthropic: Building effective agents
- Reference Model Context Protocol
- Reference OWASP Top 10 for LLM Applications