Manager Data Scientist AI

About the Applied AI Tribe

The Applied AI Tribe is talabats internal AI engine a team obsessed with turning cutting-edge AI into genuine measurable business value. We dont build AI for its own sake. Our north star is Net AI Value Delivered.

We have shipped 40 AI products across four strategic streams:

   - Product & Tech Platform - AI capabilities embedded into talabats core product and engineering platform.

   - Business Automations - streamlined workflows that save real time and money across operations.

   - Self-Serve AI Tools - empowering talabatis to build and use AI without needing to raise a ticket.

   - Training & Onboarding - upskilling the tribe so every team can participate in the AI era.

Our engineering approach is built around shipping agentic AI systems at scale with a harness-first philosophy evaluation embedded into every product iteration and production feedback loops that drive the roadmap.

The Role

As Manager Data Scientist - AI within the Applied AI Tribe you will lead a team of data scientists building and shipping the ML and generative AI systems that power decisions across talabats product and business. You are equally a people leader and a technical practitioner close enough to the work to make great technical decisions and grow your team while owning the roadmap and delivery.

Your teams work spans the full AI lifecycle: from framing ambiguous business problems through data modelling feature engineering model training deployment and production monitoring. A significant focus will be leveraging LLMs agentic systems and generative AI to automate decisions enrich data and build intelligent experiences at scale all measured against Net AI Value Delivered.
 

Whats On Your Plate

People Leadership:

   - Lead grow and retain a team of data scientists setting a high bar for technical quality fostering a culture of ownership and supporting each persons career development.

   - Partner with recruiting to attract top data science and ML talent as the tribe scales.

   - Mentor team members in ML best practices agentic system design production engineering and stakeholder communication.

   - Run effective team rituals planning design reviews retrospectives that keep the team aligned and moving with pace.

Technical Strategy & Delivery:

   - Translate ambiguous business problems into well-scoped ML and AI solutions with clear measurable success criteria tied to the tribes north star.

   - Own the teams technical roadmap: prioritise high-impact work across data enrichment business automation self-serve tooling and agentic product features.

   - Champion harness-first thinking ensure evaluation pipelines LLM observability tooling and infrastructure are in place before agent logic is built on top.

   - Embed evaluation into every product iteration: golden datasets stakeholder-aligned metrics and weekly eval jobs that guide what gets built fixed or retired.

   - Oversee the full ML lifecycle: data pipelines feature engineering model training production deployment serving and monitoring.

   - Drive adoption of LLMs and generative AI for data enrichment smart content understanding and automated decision-making at scale.

   - Design and analyse experiments (A/B and multivariate) to rigorously measure model and product impact.

   - Elevate ML and engineering standards across the team improving MLOps code quality tooling and internal learning programmes.

Cross-functional Partnership:

   - Partner with product managers and business teams to identify high-value AI opportunities and shape them into the teams roadmap.

   - Communicate clearly with senior stakeholders from problem framing through to results and recommendations.

   - Collaborate with engineering teams to understand data systems build reliable data models and ensure smooth production integration.

Qualifications :

What Did We Order

Technical Experience

   - Deep expertise in machine learning generative AI deep learning NLP recommendation systems and data mining.

   - Hands-on knowledge of ML and GenAI frameworks: Scikit-learn XGBoost LightGBM PyTorch TensorFlow Transformers and LLM fine-tuning.

   - Proficiency with the OpenAI SDK and familiarity with major LLM provider APIs for building and orchestrating production AI systems.

   - Hands-on experience with LangGraph for building stateful multi-step agentic workflows and complex agent orchestration patterns.

   - Experience with Hugging Face (Transformers Hub Inference API) for model sourcing fine-tuning and deployment.

   - Solid understanding of embeddings dense and sparse retrieval vector databases and semantic search for building RAG pipelines and similarity-based applications.

   - Strong software engineering fundamentals: clean production code data structures and algorithms ML system design.

   - Proven experience shipping and monitoring ML models in production with a solid grasp of MLOps practices.

   - Strong data and ML engineering skills: building and orchestrating data pipelines (e.g. Airflow) and robust feature engineering.

   - Excellent SQL and Python; solid statistical foundations including experiment design causal inference and predictive methods.

   - Familiarity with agentic system design (harness-first thinking LLM observability evaluation pipelines and production feedback loops) is a strong plus.

   - Experience with BigQuery and Google Cloud Platform is a plus.
 

Qualifications 

   - Bachelors degree in Engineering Computer Science or a related field. A postgraduate degree is a plus.

   - 6 years of experience in data science ML engineering and/or generative AI including shipping models to production.

   - 2 years managing or leading a data science or ML team with a track record of developing talent and delivering through others.

   - Experience building ML systems in an online consumer product environment is a strong plus.

Additional Information :

Mindset & Ways of Working

   - Curiosity over certainty: youre excited to try the new model or framework this week.

   - Ownership: you treat cost latency and business impact as your problem and your teams.

   - Bias to build: a prototype in a real workflow beats a perfect design doc every time.

   - Comfortable with ambiguity: most of what we build has no playbook yet and you help write it.

   - Keep it simple: pick the approach that delivers the most value with the least complexity and #makeithappen.
 

At talabat your success is defined by both results and behaviors. We hire for excellence in craft and for the Leadership Principles that shape how we think decide and collaborate. What you achieve matters and how you achieve it defines us. Together they move the business forward and deliver great experiences. Learn more about our Leadership Principles here.

Remote Work :

No

Employment Type :

Full-time