1FDE Mindset, Customer Discovery and Problem Framing. From "we need an AI agent" to a measured problem and a first releaselab
2Solution Architecture, System Design and Technical Scoping. Size the system, trace a request, and write the trade-offs downlab
3Software Engineering for FDEs: Python, APIs, Git and Testing. Validated requests, a client that survives failure, and tests that prove bothlab
4Enterprise Data Engineering and System Integration. Load without bad rows, sync without losing changes, keep every record traceablelab
5Cloud, Containers, CI/CD and Infrastructure as Code. Package it, run it, and put a pipeline in front of productionlab
6LLM Engineering, Model APIs and Context Design. One interface for every model, a price for every call, checked outputlab
7RAG Architecture: Ingestion, Chunking, Embeddings and Retrieval. Traceable chunks, an index that forgets old versions, answers with evidencelab
8AI Agents, Tool Calling, Workflow Orchestration and MCP. A narrow tool, fixed workflow states, and a person in charge of the moneylab
9Evaluation, Testing, Benchmarking and Model Quality. A golden set, a score per category, a readable trace and a release rulelab
10AI Security, Privacy, Guardrails and Governance. Permissions in code, attacks on your own assistant, guardrails with a defined actionlab
11Product Engineering, UX, Human-in-the-Loop and Adoption. Evidence beside the draft, every screen state, and numbers that show real uselab
12Production Deployment, SRE, Observability and Cost. Honest health checks, an error budget and the true cost of one tasklab
13Customer Delivery, Technical Communication, ROI and Productization. Results with their base, value that survives finance, and a clean handoverlab
14FDE Capstone: Design, Build, Deploy and Present. A claims-review copilot, end to end, with acceptance tests and a tenant-isolation prooflab