AI Engineering Certification
This repository holds my work for the AI Engineering Certification (AIEC v1.0) run by AI Maker Space. It is the cohort's working repo, where each session's assignments live, and I am building through it as the course progresses. The whole program is organized around a simple mantra: build, ship, and share. Every module ends with something you have actually made rather than just notes you have read.
The early sessions cover the foundations of working with large language models as a developer. This means getting comfortable with the OpenAI API and other model providers, understanding prompt engineering beyond the basics, and learning how to structure interactions so the model behaves reliably. A lot of the work here is in Jupyter notebooks, which makes it easy to iterate quickly and see exactly how a change to a prompt or a parameter affects the output.
Embeddings and retrieval-augmented generation are a major focus. I worked through how text gets turned into vectors, how vector databases store and search those embeddings, and how to wire retrieval into an LLM so it can answer questions grounded in your own documents rather than just its training data. Building a RAG pipeline end to end, from chunking and indexing to retrieval and generation, made the trade-offs concrete in a way that reading about them never did.
From there the course moves into agents. This is where the model stops being a single call and starts being a system that can reason, use tools, and take multiple steps toward a goal. I built agents that call functions, retrieve information, and decide what to do next based on intermediate results. Having done some agentic work before, I appreciated seeing how this team frames the patterns and what they consider production-ready versus a demo.
Later sessions get into fine-tuning and the operational side of running LLM applications. Fine-tuning covers when it actually makes sense versus reaching for prompting or retrieval first, and how to prepare data for it. The LLM Ops material covers evaluation, monitoring, and the unglamorous work of keeping these systems reliable once real users are hitting them. The repo also includes a TypeScript application alongside the Python notebooks, since shipping means building an actual interface people can use, not just a notebook.
The repo will keep growing as I move through the cohort. It is less a finished artifact and more a running log of the things I am building each week, with the messiness that comes with learning in public.