Senior Product Manager - Machine Learning

About the Role

Yassir's marketplace runs on decisions made millions of times a day: which driver gets which trip, what a ride should cost, when an order will arrive, which transaction looks fraudulent, who qualifies for credit. We are looking for a Product Manager who owns the machine learning systems behind those decisions and is accountable for their measurable impact on the business. This is a product role for ML-heavy, data-intensive products. You will lead discovery from data rather than from opinion, frame business problems as problems a model can actually solve, define what success means both offline and in production, and prove impact through well-designed experiments. You will work closely with data scientists, ML engineers and data engineers, and with product and operations partners across ride-hailing, delivery and financial services.

What This Role Is (and Isn't)
This role centres on predictive and decisioning ML: ranking, matching, forecasting, pricing, risk and personalisation. Experience building LLM or agentic features is welcome, but it is not a substitute for the fundamentals below. If your ML product experience is primarily integrating third-party models or APIs into user-facing features, this role is likely not the right fit.

Responsibilities

- Data-led discovery: Identify and size ML opportunities by going into the data yourself. Diagnose where the marketplace is losing value (supply-demand imbalance, cancellations, ETA error, fraud losses, default rates) and translate it into a prioritised, quantified problem backlog.

- Problem framing: Turn business problems into well-posed ML problems: define the prediction target, the decision it informs, the unit of analysis, label availability and quality, and the cost of different error types. Know when a problem does not need ML and a rule or heuristic will do.

- Metrics and success criteria: Define the chain from model metrics (e.g. precision/recall, calibration, MAE) to product and business metrics, and own guardrail metrics. Understand why offline gains often fail to translate online, and plan for it.

- Experimentation and causal inference: Design and interpret controlled experiments, including in two-sided marketplace settings where interference makes standard A/B tests unreliable (switchback, geo or cluster-randomised designs). Reason about statistical power, novelty effects and heterogeneous impact across cities and segments. Distinguish correlation from causation, and know which quasi-experimental methods to use when randomisation isn't possible.

- ML product lifecycle: Own models from framing through data requirements, baseline, iteration, launch, monitoring and retirement. Partner with engineering on rollout strategy, model monitoring, drift detection, retraining cadence and failure modes. Treat a model in production as a product that degrades if unattended.

- Roadmap and trade-offs: Own the ML product roadmap across domains and countries. Make explicit trade-offs between accuracy, latency, cost, interpretability, fairness and regulatory requirements, especially in credit and financial services.

- Stakeholder alignment and visibility: Make the impact of ML work legible to non-technical leadership. Communicate results, including null and negative results, with rigour and clarity. Align product, operations, risk and marketing partners on shared objectives.

- Leadership and culture: Raise the bar on how the organisation makes decisions with data. Mentor peers, build a culture of feedback and trust, and invest in your own growth and that of those around you.

Requirements

- 4+ years of product management experience, including at least 2 years owning ML-driven products that run in production and influence core business decisions (e.g. pricing, matching, ranking, forecasting, fraud, credit risk, recommendations).

- Hands-on fluency with data: you can write SQL, explore data independently and challenge an analysis without waiting for someone else to run it.

- Demonstrated experience designing, running and interpreting controlled experiments, and a solid working understanding of causal inference.

- Strong understanding of core ML concepts: supervised learning, evaluation metrics and their trade-offs, overfitting and leakage, bias, calibration, and the relationship between offline and online performance.

- A track record of shipping ML products where you can clearly articulate the problem framing, the metrics chosen, the experiment design and the measured business impact, including what went wrong.

- Experience working with distributed or remote teams across multiple markets.

- Experience in marketplaces, mobility, on-demand delivery or fintech is a strong plus.

- BSc/MSc in Engineering, Computer Science, Statistics, Data Science or a related quantitative field.