Agents that do the work, not demos that do nothing.
I'm Anas Nisar, a lead engineer who designs and ships production-grade AI agents — the kind that take real workflows off your team's plate and keep running when it matters.
How I work
Find the work worth automating
Not everything should be an agent. I help you spot the few workflows where autonomy actually pays for itself — and talk you out of the ones that don't.
Make it survive production
Evals, guardrails, fallbacks, and real integration into your systems — so the agent still works on day ninety, not just in the demo that won the meeting.
Hand it over cleanly
Monitoring, documentation, and a team that understands what's running. You get a system you can trust, not a black box you're afraid to touch.
Writing
Keeping agents fast: lean state, direct tools, and the data-card trick
An agent that's correct but slow loses to one that's slightly worse and snappy. Most latency in agent systems is self-inflicted — here's where it hides.
You can't fix what you can't trace: observability for multi-agent systems
When an orchestrator hands off to a sub-agent that calls a tool that calls a model, 'it gave a bad answer' isn't a bug report. Here's the instrumentation that turns it into one.
Memory is not context: how agents remember without lying
Conflating long-term memory with live state is the fastest way to make an agent confidently wrong. The fix is a hard line between what the agent recalls and what it looks up.
Facts, affordances, and pertinence: who decides what an agent does — code or the model
Two failure modes haunt every agent: it makes up things that aren't true, and it nags about things at the wrong moment. Both come from putting the wrong decision in the wrong place.
About
I've spent seven years building scalable production systems — most recently leading the engineering for a consumer AI product, where I design the agents, the guardrails that let them act safely, and the behavioural-data pipelines that feed them. I care about the unglamorous part of this field: getting agents to behave reliably once they leave the lab and meet real users, real data, and real consequences.
Now I help businesses do the same — turning the manual, repetitive, expensive parts of their operations into agents that run quietly in the background and hold up under load.
If that's the kind of problem you're sitting on, let's talk.
Have a workflow that should run itself?
Tell me what's eating your team's time. If it's a fit for an agent, I'll tell you how I'd build it — and if it isn't, I'll tell you that too.
anasnisar980@gmail.com →