Infrastructure$349/ seat
Edge-first architecture
Design globally distributed apps that are fast everywhere and fail gracefully.
Workshop · June 9
Production LLM apps
Build, evaluate and ship a retrieval-augmented assistant that holds up in production.
Overview
In one intensive day you'll build a retrieval-augmented assistant from scratch, then spend the afternoon on the parts demos skip: evaluation, guardrails, cost control and observability.
You'll leave with a working repository, an evaluation harness you can reuse at work and a checklist for taking LLM features to production.
Highlights
Retrieval that works
Chunking, embeddings and hybrid search, tested on real data.
Evaluation harness
Automated tests for answers, not just code.
Guardrails
Prompt injection defences and safe fallbacks.
Cost control
Caching and routing to cut token spend by 60%.
What's included
- Laptop with Python 3.11+
- Free API credits provided
- Starter repository
- Lunch and coffee
- Certificate of completion
- Recording access for 90 days
Agenda
How retrieval-augmented generation works, and where it breaks.
Ingest documents, embed them and wire up search and generation.
Write automated evaluations and find the failure cases.
Guardrails, caching, monitoring and a production checklist.
More workshops
You might also like
Infrastructure$349/ seat
Design$299/ seat
Design systems at scale
Tokens, components and governance for systems used by hundreds of teams.
Security$299/ seat
Threat modelling for product teams
A practical, no-jargon method to find security issues before attackers do.