Agentic AI Engineer

Agentic AI Engineer

Join an early-stage team building LLM-powered agents into the core of an MVP. You will work shoulder-to-shoulder with our Senior AI Engineer to turn experimental agentic workflows into reliable features that early users actually depend on. This is a hands-on role for someone who wants to learn agent design, tool integration, and evaluation the right way — by shipping and reading traces, not by watching demos.

Why this role exists

Our data lives in fragmented sources and our early users need intelligent interactions on top of it — fast. We need someone who can help us quickly build, test, and integrate our first LLM-powered agents into the MVP so we can start delivering value now, while laying groundwork (clean tool interfaces, basic evals, reproducible dev environments) that keeps the system honest as we scale.

What you'll do

Agent development. Partner with the Senior AI Engineer to design, build, and test LLM-powered agents in Python using LangChain. Start with the simplest control flow that works — prompt chains and routers — and escalate to agentic loops only when the task genuinely needs it.

Data and API integration. Connect agents to our fragmented data sources and external APIs. Write tight tool schemas, handle errors as structured feedback the agent can recover from, and make sure tool calls are reliable before worrying about anything cleverer.

Reliability and debugging. Read agent traces, categorize failures, and help build the small evaluation loop that tells us whether a prompt or tool change actually helped. Troubleshoot response accuracy issues and tune workflows based on what the logs show, not what the demo feels like.

Engineering hygiene. Write clean, well-documented Python. Manage local development environments with Docker so the team can reproduce runs. Treat prompts as versioned artifacts, not throwaway strings.

What we're looking for

Solid Python fundamentals — you can build, test, and debug non-trivial code.

Hands-on experience with an LLM framework like LangChain or LlamaIndex (projects, internships, or side work all count).

Comfort integrating REST APIs and querying data with basic SQL.

Familiarity with Docker for local development.

A proactive, self-starting mindset — you ask questions, read traces, and iterate quickly.

Curiosity about why agents fail, not just when they succeed.

Nice to have

LlamaIndex or other retrieval frameworks.

Exposure to evaluation practices for LLM systems (golden sets, LLM-as-judge).

Prior work on tool/function-calling design or structured output.

What success looks like

Within 90 days you will have deployed a fully tested agentic workflow that interacts with our fragmented data sources, with reliable response rates and successful API integrations serving our early users. Along the way you will have built the debugging and evaluation habits that let us keep improving the agent without breaking it.

Logistics

Fully remote from Egypt. Full-time role, EGP 20,000-30,000 per month. Reports to the Senior AI Engineer, working within a small engineering team of 2-5. Application deadline: 16 July 2026.

Benefits

Fully remote work

Annual learning budget for books, courses, and conferences

Direct mentorship from a Senior AI Engineer on production LLM systems