Senior Data Scientist - at Kamino Retail - Full-remote (France)

📋 Description

• • Architect the first deep-learning algorithms for Kamino Retail, a fast-growing SaaS platform that processes billions of logs daily and is already trusted by leading retailers across Europe. You will start from a blank canvas—no legacy code, no pre-conceived constraints—giving you the rare freedom to define the technical direction of our entire data-science stack.
• • Translate high-impact business questions—“Which ad placement will maximize incremental revenue?” “How do we predict cart abandonment in real time?”—into rigorous machine-learning problems. You will own the full lifecycle: framing the hypothesis, selecting features, training models, and measuring uplift in live A/B tests.
• • Design and implement always-on training pipelines that run on our Kubernetes cluster and feed a high-performance ClickHouse data mart. You will decide how often models retrain, which drift-detection thresholds trigger redeployments, and how to version datasets so that every experiment is reproducible.
• • Deploy models behind low-latency REST and gRPC services that must respond in milliseconds while handling tens of thousands of requests per second. You will choose the right combination of TensorFlow Serving, ONNX, or custom Python micro-services, and you will monitor latency, throughput, and error rates with Prometheus and Grafana.
• • Establish MLOps best practices: CI/CD for training code, automated unit tests for data transformations, feature-store governance, and model cards that document fairness and bias metrics. Your standards will become the template that every future data scientist at Kamino follows.
• • Collaborate daily with a dedicated data engineer who is building streaming ingestion, feature pipelines, and experiment-tracking infrastructure. Together you will decide on schema evolution, partitioning strategies, and cost-efficient resource allocation on GCP.
• • Influence product strategy by presenting findings to the CPO and retail clients. Your dashboards and slide decks will turn complex model outputs into clear recommendations that merchandising teams can act on the same day.
• • Mentor junior analysts and data scientists through pair programming, design reviews, and weekly learning sessions. You will foster a culture of scientific rigor, curiosity, and customer obsession.
• • Stay ahead of the retail-media curve: test new transformer architectures for sequential-basket prediction, explore reinforcement learning for dynamic pricing, and publish internal white-papers that keep Kamino at the cutting edge.
• • Enjoy full-remote flexibility within France, asynchronous communication, and the autonomy to organize your day around deep-work blocks and customer meetings. Your impact is measured by shipped models and business KPIs, not by hours online.