AI Engineer, Knowledge & Retrieval Systems

Company Description I am building a new AI product that helps organisations get far more value from what they know. We’re looking for a hands-on engineer to help design and build it from the ground up. This is an early-stage role. You’ll work directly with the founder, shape the architecture, and see your work used by real companies. What you’ll work on: * Designing how knowledge is structured, stored and retrieved for AI systems * Building pipelines that turn documents, data and conversations into structured, usable knowledge * Connecting AI agents to that knowledge through APIs and the Model Context Protocol (MCP) * Building evaluation frameworks that measure whether the AI actually performs better * Turning prototypes into secure, reliable products Skills we’re looking for: * Building applications with large language models: prompting, tool use, agents and structured outputs * Retrieval-augmented generation (RAG): vector and hybrid search, chunking, re-ranking * Knowledge modelling: schemas, knowledge graphs or ontologies (e.g. Postgres, Neo4j) * Evaluating LLM systems: test sets, rubrics, LLM-as-judge, comparing approaches * Backend development: Python, APIs (e.g. FastAPI), Docker, AWS, Azure or GCP * Data pipelines and integrations with business systems (CRM, ticketing, documents) * Security basics for business software: access control, multi-tenancy, data privacy Nice to have: * Experience with agent memory frameworks (e.g. Mem0, Letta, Zep) or MCP servers * Fine-tuning open-weight models * Experience building for enterprise or B2B customers * Arabic language skills * Interest in research and contributing to published work What we value: * You ship fast and learn by building * You’re comfortable with ambiguity and early-stage work * You communicate clearly and take ownership * You use AI coding tools to move faster Details: 📍 Remote, with some overlap with Gulf working hours (GMT+4) 🕒 Project based, Freelance or contract to start, with potential for a long-term role or partnership 🔒 All work is covered by an NDA and IP assignment How to apply: Include in your response: 1. Your GitHub, portfolio, or a project you’re proud of 2. Two or three sentences on an AI system you’ve built that used retrieval or memory, and what you learned from it 3. Your availability and expected rate I welcome applicants at all experience levels. Show me what you’ve built!