Seven modules. One running enterprise project. Production-ready AI engineering literacy.
The Common AI Core every engineer completes before specialising. You build one Northgate Bank assistant and evolve it, module by module, from a single model call into a production-style AI application.
One Northgate Bank assistant, evolved across seven modules into a production-style AI application.
Can explain how LLMs behave and choose appropriately between prompting, RAG, fine-tuning, workflow, or agent
Can design a reliable prompt/context flow and diagnose whether a bad response is caused by instructions, context, or the model
Can define expected behavior, build a small eval set, measure quality, and compare two versions objectively
Can explain and build the basic ingestion → retrieval → context → generation pipeline and diagnose retrieval failures
Can build a basic tool-using agent/workflow and explain state, memory, approvals, permissions, and agent failure modes
Can trace an AI request, identify latency/cost/failure issues, and explain basic caching, retries, rate limits, and reliability patterns
Can identify common AI risks, design basic controls, and explain data boundaries, authorization, auditability, and governance expectations