AI Engineer · Python & Agentic Systems
Build and ship agentic systems in Python alongside a lead — tools, evals, and the production plumbing that keeps them honest.
What this seat actually is.
The same work as the lead role, one rung down in scope. You build agentic systems in Python and take real ownership of components, while a lead holds the overall architecture and the client conversation.
You are expected to be self-directed inside a defined design: pick up an agent surface, build it properly, write the evals, take it to production, and hold it there. What you are not expected to do yet is set the architecture for a whole engagement alone — you will spend a lot of time in review with someone who does, and that is the point.
This is a mid-senior seat, not a junior one. We do not hire juniors, and there is nobody here whose job is to unblock you daily.
Fully remote from the UAE, async-first.
What you own.
- ·Build agent components in production Python: tools, orchestration steps, structured output handling, retries, and guardrails.
- ·Write and maintain evals for the surfaces you own — golden sets, regressions, and the metric that would catch a quiet degradation.
- ·Implement retrieval and context pipelines: ingestion, chunking, indexing, and the plumbing that keeps them fresh.
- ·Integrate agents with real client systems — APIs, queues, databases, and the authentication and rate-limit reality around them.
- ·Instrument what you ship: latency, cost, token usage, tool-call success, and error taxonomy.
- ·Take your own work to production — container, pipeline, deploy, and the dashboards that prove it is healthy.
- ·Contribute to architecture in writing and argue for a better design when you see one.
- ·Take part in the on-call rotation for the systems you helped build.
Every line here is real.
- ·Solid mid-senior Python engineering with production services shipped and operated
- ·Agentic framework experience is required — you have built multi-step tool-using agents, not only chat wrappers
- ·Working grasp of solution architecture: you can read a system design, find the weak joint, and propose the fix
- ·Practical LLM experience: prompting, tool schemas, structured output, context limits, and the usual failure modes
- ·Evals as a habit — you measure a change rather than eyeball it
- ·Comfortable with APIs, async work, queues, and SQL
- ·Familiar with Docker and shipping through CI to a real environment
- ·Clear written English; you can explain a design decision without a call
- ·Self-directed inside a defined architecture and honest about what you do not know
- ·Based in the UAE and able to work a UAE-anchored schedule
- ·Hands-on with Claude or Claude Code, or curiosity about MCP servers
- ·RAG experience, including where it did not work and why
- ·Kubernetes exposure
- ·Front-end capability for internal tools and eval dashboards
- ·Arabic in addition to English
- ·A shipped side project you can talk about honestly
One email, one reply within 5 days.
The button below fills in the subject line, so your application reaches the right principal.
Open email →Subject: Application - AI Engineer · Python & Agentic Systems
- CVPDF preferred. One or two pages is plenty — we read all of it.
- WhyOne paragraph: why this role, and something you shipped end to end.
- LinksAnything public you are proud of. Optional, never penalised.