LLMOps Assistant

An end-to-end LLMOps pipeline on Google Cloud — BigQuery → Vertex AI tuning → Kubeflow automation → deployment — that fine-tunes a foundation model to answer machine learning / deep learning engineering questions, using real Stack Overflow data, then serves it behind an API with prompt-template consistency and safety checks.

Applies the LLMOps lifecycle covered in DeepLearning.AI's LLMOps Pipeline course to a scoped, original use case: an assistant specialized in ML/AI engineering questions rather than a general coding assistant — a narrower, more defensible domain that's closer to what building this kind of system looks like on the job.