About the role
This role builds production agentic systems in Python. You'll own real components end to end, from implementation and evaluation to deployment and operational support, while working within an architecture defined by the Lead AI Engineer.
It is a self-directed engineering role inside a well-defined technical direction. You'll build production-ready capabilities, measure their quality, ship them, and keep them healthy over time.
What you'll do
- Build agent components in production Python: tools, orchestration steps, guardrails.
- Write and maintain evals for the surfaces you own.
- Implement retrieval and context pipelines.
- Integrate agents with real client systems: APIs, queues, databases.
- Instrument latency, cost, and tool-call success.
- Take your own work to production.
- Join the on-call rotation for what you help build.
Must have
- Strong Python engineering experience delivering production services.
- Experience building multi-step, tool-using agents, not simple chat wrappers.
- A working grasp of solution architecture: you can identify weak points in a design and contribute to improving it.
- Practical LLM experience: prompting, tool schemas, structured outputs, and function calling.
- Evals as a habit: you measure changes rather than relying on intuition.
- Comfortable working with APIs, asynchronous programming, queues, and SQL.
- Familiar with Docker and deploying through CI/CD pipelines.
- Self-directed, accountable, and honest about what you don't know.
Nice to have
- Hands-on experience with the Model Context Protocol (MCP).
- Experience building Retrieval-Augmented Generation (RAG) systems.
- Exposure to Kubernetes.
- Ability to build lightweight internal tools and dashboards.
- Experience working with enterprise or government systems.
How you work
We value these qualities as highly as your technical skills.
- You care about shipping reliable software, not impressive demos.
- You write code that other engineers can understand and maintain.
- You measure quality instead of relying on intuition.
- You ask questions early when something doesn't make sense.
- You communicate technical ideas clearly to engineers and non-engineers alike.
- You care about latency, reliability, and cost, not just model quality.
- You take ownership of the systems you build after they reach production.
- You work at a sustainable pace and collaborate well with the rest of the team.
- Professional working proficiency in English and Arabic. You'll communicate directly with client stakeholders in both languages, so both are required for this role.
Stack
- Python
- Claude
- FastAPI
- Postgres + pgvector
- Docker
- CI/CD
- AWS