Production AI & Context · preview
Build AI systems that ship
Six-week program — context engineering, prompting patterns, multi-step agent orchestration, and tmux for serious terminal workflows. Includes 1:1 live lectures every week on Zoom and one-on-one Swift Chat with your instructor.
What you'll work on
- → AI tools — choosing and combining models, APIs, and local tooling for real tasks.
- → Context engineering — shaping prompts, memory, and retrieval so systems behave reliably.
- → Prompting patterns — instruction, constraints, evaluation, and iteration loops.
- → Orchestrating agents — multi-step flows, delegation, guardrails, and failure handling.
- → tmux — terminal multiplexing for long-running jobs and reproducible CLI workflows.
Prerequisites
You should already have programming foundations, comfort with DevOps-style workflows (shell, environments, deploy basics), and experience integrating working APIs. This is not an entry-level coding course.
Inside the workspace (unlocks after enrollment)
Week 1 · Toolchain
Pick a model, set up local tooling, scaffold your first context-aware loop.
Week 2 · Context engineering
Memory, retrieval, evaluation harnesses — make outputs predictable.
Week 3 · Prompting at depth
Patterns, constraints, structured outputs, edge case handling.
Week 4 · Multi-agent orchestration
Delegation, hand-offs, guardrails, deterministic failure modes.
Week 5 · tmux + production CLI
Long-running jobs, panes, reproducible terminal workflows.
Week 6 · Ship a real system
Capstone — design, build, evaluate, deploy.