Muhammad Junaid

I'm Junaid — an AI Engineer from Kotli, Azad Kashmir, Pakistan, building and shipping real products for clients worldwide, remotely. My work sits at the intersection of applied machine learning and full-stack engineering: I design GenAI and LLM systems (RAG pipelines, tool-using agents, fine-tuned models) and build the production software around them — the APIs, interfaces, and infrastructure that turn a model into something people actually use.

A lot of AI work stops at the demo. My focus is the harder 80% after that: making systems accurate, fast, observable, and maintainable. That means caring about retrieval quality and grounding as much as prompts, treating latency and cost as features, and writing code that a team can live with. I've worked across the stack — Python and PyTorch on the ML side, TypeScript, React, Astro and Node on the product side, with Docker, cloud, and Postgres holding it together.

I also write. On this blog I share what I learn building with AI — practical breakdowns of RAG, fine-tuning, LLM agents and the engineering decisions that separate a prototype from a product. If you're working on something interesting — a product, a role, or a hard problem — reach me via the contact form or hello@junaidai.com.