About
I came to AI engineering through software engineering, and it shows.
I’m Hussain Ahmad, an AI Software Engineer based in Faisalabad, Pakistan. I build agent systems, retrieval-augmented applications and the full-stack software that has to support them.
My background is in building conventional systems — APIs, access control, queues, realtime delivery, test suites, deployment pipelines. That turned out to be the useful preparation. Most of what makes an AI feature trustworthy is not model work: it is validation at the boundary, work moved off the request path, provenance carried with an answer, and a way to tell whether last week’s change made anything better.
So I treat a model as one component inside a system that has to keep its promises. I’m interested in where that component should be trusted, where it should be constrained, and where it should hand back to a person.
I work in the open. Everything on this site links to source, and where a result has not been measured I say so rather than filling the gap with a number.
- Role
- AI Software Engineer
- Based in
- Faisalabad, Pakistan · PKT (UTC+5)
- Focus
- AI agents, RAG, LLM applications, full-stack systems
What I look for
The work I want to be doing
Being specific about this saves everyone time, including me.
Problems with a real cost of being wrong
Systems where an incorrect answer has a consequence are more interesting than ones where it does not, because they force the design questions that matter: confidence, escalation, auditability.
Work that reaches production
Prototypes are useful and I build them, but the engineering I care about starts at the point where other people begin to depend on the thing.
Teams that write things down
Architecture decisions, trade-offs and the reasons behind them. It is the difference between a codebase that can be changed and one that can only be added to.
Currently exploring
What I'm working through right now
Open questions I'm spending time on, rather than a list of things I already know.
Agentic AI
Multi-step agents that plan, call tools and know when to hand back to a person.
AI evaluation
Making prompt and retrieval changes measurable against a fixed question set instead of judged by feel.
Production RAG
Ingestion freshness, retrieval quality and citation integrity at the point where a corpus stops being small.
Long-term agent memory
What an agent should keep between sessions, and what it should be made to forget.
Multi-agent architectures
When splitting work across specialised agents beats one capable agent, and when it only adds coordination cost.
AI infrastructure
Queues, workers, streaming and observability — the layer that decides whether an AI feature survives contact with users.
Contact
Have an AI product or workflow worth building?
Send me the problem — not the spec. If it’s a fit I’ll tell you how I’d approach it; if it isn’t, I’ll say so.
Faisalabad, Pakistan · PKT (UTC+5) · Working remotely