Advanced

Advanced LLM Engineering

Production-grade techniques for engineers building serious AI systems.

Programme Overview

A deep, technical programme for software engineers and ML practitioners building real LLM-powered applications. Covers retrieval-augmented generation, agentic architectures, evaluation pipelines, cost and latency optimisation, and the operational realities of running LLM systems at scale.

What You'll Learn

Architect production-grade RAG systems
Design and evaluate multi-step agentic workflows
Build robust evaluation pipelines for LLM applications
Optimise for cost, latency, and reliability in production
Implement guardrails and failure handling at scale

Who It's For

Software engineers shipping LLM-powered features
ML engineers moving into generative AI
Technical architects evaluating LLM platforms

Prerequisites

Proficient in Python
Prior exposure to APIs and backend development
Basic familiarity with LLM prompting

Curriculum

- Structured outputs and tool use
- Function calling patterns
- Prompt chains and composition
- Embedding models and vector stores
- Chunking, reranking, and retrieval quality
- Hybrid search and query rewriting
- Tool-using agents and loop design
- Multi-agent coordination patterns
- Failure modes and recovery
- Offline evals and regression suites
- Online evaluation and user feedback loops
- Cost, latency, and observability