Decisioning sits at the core of Markin. Every signal we capture eventually feeds a model that ranks candidate actions per customer, per moment. You will design, train and productionize the models that make those calls — with hard latency budgets and real revenue at stake.
What you'll do
Design and ship contextual bandit and uplift models running online at enterprise scale.
Own the full lifecycle: offline evaluation, shadow deployment, ramp, monitoring, retraining.
Partner with forward-deployed engineers to translate business objectives into reward functions.
Push the state of the art on counterfactual evaluation and off-policy learning.
What we're looking for
5+ years shipping ML in production, ideally in ranking, recsys, ads or decisioning.
Deep grasp of causal inference, uplift modelling or reinforcement learning.
Fluent in Python, PyTorch and modern data tooling.
Comfortable owning latency, cost and quality trade-offs end to end.
Nice to have
Published research at NeurIPS, ICML, KDD, RecSys or similar.
Experience with real-time feature stores and online serving.
Apply
Senior ML Engineer, Decisioning.
Tell us about you. We read every application and reply within a week.