I wrangle messy, real-world data into products your users actually love—from classical ML to agentic systems—delivered with calm, drama-free MLOps.
About Me
ML-Data Scientist and Tech Lead with 10+ years of experience building and shipping machine learning at scale. I train models, design multi-agent systems, and architect MLOps pipelines—the whole path from prototype to production. My engineering background means I care as much about how a system runs in prod as how it performs on a benchmark. I lead teams that build clever, reliable things and get them out the door.
About This Blog
The Overfitting Club has two sides. Tutorials are detailed, from-scratch guides where I break down topics that matter for my work—deep learning, MLOps, edge AI, and whatever comes next. Every tutorial is a runnable notebook with real code, real training, and real outputs. The philosophy is learn by building, not by reading summaries. Thoughts are essays and opinions about the Data Science, ML, and AI world from a practitioner’s perspective.
Writing is how I learn and relearn. If it helps someone else along the way, even better.
Current Interests
Deep Learning on the Edge: Quantization, pruning, distillation, and other tiny sorcery to run big brains on small chips (Federated Learning, Differential Privacy, etc.).
Tiny agents on device: Local LLMs with manners: tool calling, confident with local data, and calling the cloud only when it truly matters.
Why? Because GPUs don’t grow on trees (tragic), so we build clever, resource-savvy systems that do more with less.
How I Write
Essays and thoughts are entirely my own. For the tutorial series, Claude assists with code iteration, copyediting, and diagram refinement—the same way I’d use any good development tool. The ideas, structure, and editorial decisions are always mine; the AI accelerates the process.