Skip to content
PreReqs

AI Engineering

Skill matrix.

Use this as a prioritization board: what each skill area requires at production level, and how to prepare it.

Skill areaWhat "production-ready" requiresPriorityHow to prepare
Enterprise Gen AI architectureDefine a foundation architecture and technical roadmap for a Gen AI platform.P0Prepare a 2-minute reference architecture and a 6-month roadmap narrative.
RAG architectureDesign and reason about production retrieval-augmented generation systems.P0Study ingestion, chunking, embeddings, hybrid retrieval, reranking, citations, and evaluation.
LLM fine-tuningKnow when and how to fine-tune versus retrieve versus prompt.P1Prepare a RAG-vs-fine-tuning-vs-prompting decision framework with concrete tradeoffs.
LLMOpsOperate prompts, models, and evidence as a running production system.P0Learn prompt/model registries, eval harnesses, drift detection, rollback, and cost/latency monitoring.
EvaluationDefine and measure quality before and after shipping.P0Practice designing a golden dataset and a scoring rubric for a given use case on the spot.
Guardrails and safetyDesign input/output controls that hold under adversarial and edge-case input.P0Study prompt-injection defense, output scanning, and human-in-the-loop gates for irreversible actions.
Agentic systemsBound autonomy, tool access, and state in multi-step agent workflows.P1Study control-loop termination conditions, tool-permission scoping, and failure-mode design.
Security, privacy, governanceApply access control, PII handling, and regulatory-risk thinking to AI systems.P1Study risk tiering, audit logging, and the difference between accepted risk and a hard blocker.
Cost architectureReason about unit economics and budget planning for a Gen AI platform.P2Practice building a Bill of Materials and unit-economics model for a sample system.