How I Built an End-to-End MLOps Pipeline: MLflow + FastAPI + Kubernetes
Why most ML portfolios stop at the notebook and why that's wrong. A complete walkthrough of the architecture, from experiment tracking to production monitoring.
MLOps in 2025: The Stack I Actually Use (and Why)
There are 47 MLOps tools. Here's the opinionated stack I've settled on — MLflow, Prefect, FastAPI, Docker, Kubernetes, Prometheus, Grafana, and GitHub Actions — and why each one earned its place.
Building a RAG Pipeline from Scratch — What Actually Matters
Moving beyond the toy RAG demo. How to build a production-grade retrieval-augmented generation system that handles real documents and real queries reliably.
FastAPI in Production: What the Docs Don't Tell You
After deploying multiple FastAPI services to production — here's the non-obvious stuff: middleware patterns, structured logging, Prometheus integration, and graceful shutdown.