Company Description
Numu, meaning “growth” in Arabic, is an AI-managed growth function for restaurants and retailers, focused initially on online delivery. Rather than offering software for teams to operate, Numu provides an end-to-end AI function that actively manages and scales delivery channels on platforms such as Deliveroo, Talabat, Careem, Noon and other GCC services. Proprietary AI agents perform continuous diagnostics, anomaly detection, refund & review auditing, menu engineering, and campaign management across locations and platforms to grow revenue and profits. A lean operations layer oversees the system, handling exceptions and ensuring reliability. Based in Dubai with operations across the GCC and the Middle East, Numu enables brands to grow delivery channels without building internal management teams.
Role Description NIA is NUMU’s intelligence analyst and agent orchestrator: the AI system that turns delivery-platform data from many brands into diagnostics, decisions and actions for our clients. This role builds the agents inside it, the context they reason over, and the harnesses that prove their answers are right before a client sees them. You will work directly with the founder.
What you’ll do:
- Design and ship AI agents end to end, from the reasoning loop and tools through to the client deliverables they produce.
- Build the context layer agents reason over: brand context, platform context, and market benchmarks, structured so agents can use it reliably at scale.
- Own the evaluation harness: automated tests that grade agent output against real data, including adversarial cases, so reliability is measured rather than assumed.
- Design the governance production agents need: memory governance, prompt-injection defenses and strict data isolation.
- Develop the engineering processes, tooling and internal platforms that let a very small team ship agents quickly and safely, using AI coding tools throughout
Qualifications
- Strong foundation in Computer Science and Software Development, including data structures, algorithms, and production-grade coding practices.
- Solid understanding of Neural Networks and Pattern Recognition, with experience designing, training, and evaluating modern machine learning models.
- Hands-on experience with Natural Language Processing (NLP), especially applying NLP techniques to real-world business and operations data.
- Proficiency with ML/AI frameworks and tools (e.g., Python, PyTorch or TensorFlow, scikit-learn, modern NLP libraries) and working with cloud or container-based environments.
- Bachelor’s or Master’s degree in Computer Science, Engineering, Data Science, or a related field, or equivalent practical experience.
- Ability to work independently in a remote, distributed environment, communicate clearly with cross-functional stakeholders, and deliver reliable solutions on schedule.
- Experience with data engineering workflows, RESTful APIs, and MLOps practices (CI/CD, monitoring, model deployment) is highly beneficial.
- Exposure to e-commerce, delivery platforms, or retail/restaurant analytics in the GCC or Middle East region is a plus.
- Huge sense of ownership, from scoping the task to making sure it's closed and delivered.
What success looks like
- You reach for AI first. You build with AI coding agents every day and use them to multiply your output; hand-writing what an agent could draft is the exception.
- You give product feedback. NIA is a 0-to-1 product. You say what is broken or missing and fix it rather than working around it.
- You take on new tasks. We are an extremely lean team and fulfil new work in-house before we hire for it. You see that as an opportunity, not scope creep.
- Your agents can be trusted. Their answers hold up against real data, and your evals catch regressions before a client does
