Senior ML Engineer (IC4) - Yield Optimization

Responsibilities

• Own ML systems end to end across the Yield Optimization portfolio (Demand Shaping, Floor Optimization and Bid Optimization), from the business problem and model formulation through evaluation, deployment and running the model in production.
• Build and operate real-time inference services that score bid requests inside the auction, within millisecond latency budgets and at billions of requests a day.
• Scale training and batch workflows on GCP (Airflow, Spark, BigQuery, Vertex AI) so models move from experiment to production quickly and reproducibly.
• Build feature pipelines and the feature store, and keep features consistent between offline training and online serving.
• Run model serving on Kubernetes (GKE) and NVIDIA Triton, including packaging, rollout, autoscaling and cost.
• Set up CI/CD, versioning for data, code and models, and monitoring and alerting for models in production. Take part in the on-call rotation.
• Work with data scientists on modelling choices and A/B tests, and judge what can be served and maintained at scale.
• Partner with ad-server engineering, data engineering and Product to integrate models into the serving stack and plan the technical roadmap.
• Write design docs and make architectural trade-offs between latency, throughput, cost and technical debt.
• Lead projects across teams and mentor engineers and data scientists to raise the team's engineering standards.

Required Skills

• 5+ years of software engineering experience, including at least 3 years building and running ML systems in production.
• Strong Python and SQL. Experience with a compiled language (Go, Java, Scala or C++) for performance-critical services.
• Solid software engineering practice: testing, code review, API design (gRPC or REST) and distributed systems.
• Hands-on MLOps with Docker, Kubernetes, CI/CD and infrastructure as code (Terraform).
• Experience with GCP data and ML services (BigQuery, GKE, Vertex AI, Dataproc), or the equivalent on AWS or Azure.
• Experience serving models at low latency and high throughput (Triton, TensorFlow Serving or similar), and with optimization techniques such as quantization and request batching.
• Enough ML depth to work as a peer with data scientists: PyTorch, TensorFlow, scikit-learn or gradient-boosting libraries, and how models are trained and evaluated.
• Large-scale data processing with Spark/PySpark and workflow orchestration with Airflow.
• Experience owning production systems, including monitoring, incident response and on-call.
• Clear communication of technical trade-offs to engineers, data scientists and product managers.
• Bachelor's or Master's degree in Computer Science, Software Engineering or a related quantitative field, or equivalent experience.

Nice to Have

• • AdTech experience: programmatic auctions, real-time bidding and supply/demand matching.
  • Feature store experience (Feast, Tecton or similar) with an online store such as Redis.
  • Stream processing with Kafka or Flink.