Flinks

• Own and evolve the data platform - the BigQuery warehouse, dbt transformation layers, Airflow / Cloud Composer orchestration and Pub/Sub ingestion that feed every model and metric. • Build and operate the ML platform - training pipelines (Kubeflow on Vertex AI), model serving (FastAPI behind Vertex endpoints), CI/CD, containerization and typed contracts. • Take operational ownership of model-serving infrastructure so reliability isn't carried by the data scientists alone. • Harden and standardize the data models the business depends on - improving schemas, fixing data-quality issues and establishing trustworthy source-of-truth feeds. • Establish data governance and observability - bring data that lives outside the warehouse under proper governance and build operational metrics for products that don't yet have them. • Standardize how data engineering is done across product lines - patterns, tooling and pipelines other teams can adopt. • Partner across data science, backend and product on the producer to consumer contract (models produced by data science, consumed/aggregated downstream, surfaced to clients).

• Own ML models end to end - frame the problem, write the design/RFC, build and train the model, ship it to a served endpoint and monitor its quality in production. • Build your model's training pipeline and package it for serving, deploying onto the shared platform Data Engineering builds and operates. • Own model quality, not the platform - you watch drift and performance and decide when to retrain. • Evaluate rigorously - experimental design, statistical validation, drift detection and retraining, champion-challenger evaluation and promotion. • Partner across the org - your model outputs feed the attributes/enrichment layer, payments risk, dashboards and client integrations, you'll collaborate with Data Engineering, backend, product and QA on contracts, deployment and rollout. • Move fast with AI-assisted development - we use it to accelerate implementation and experimentation, the highest-leverage contribution in this role comes from strong problem framing, system design, evaluation rigor and clear technical specifications.