AI Data Engineer

Own data pipelines end-to-end: Design, build, and operate batch/streaming ETL/ELT (positions, trades, prices, fundamentals, benchmarks, corporate actions, CRM, clickstream).

ML feature platform: Stand up and maintain offline/online features (e.g., client risk signals, asset factors, churn/upsell, RIA productivity), versioning, and lineage.

Model training & MLOps: Productionize models (forecasting, personalisation, anomaly/fraud detection, NLP document processing) with automated training, CI/CD, canary deploys, and drift/quality monitoring.

Data quality & governance: Implement contracts, unit tests, expectations, schema registry, lineage, catalogue, PII tokenisation, and role-based access.

Serving & APIs: Expose model/feature services with low-latency endpoints and scalable infra; partner with backend to integrate into advisor & client apps.

Cross-functional impact: Collaborate with Quant/Research, Product, and Compliance to translate business problems into robust data/ML solutions and measurable outcomes.

5+ years in data engineering/ML engineering (or equivalent), including owning production pipelines and shipping ML into production .

Expert in Python and one of SQL/Scala ; strong with Spark/PySpark and a workflow orchestrator (Airflow/Prefect/Dagster).

Hands-on with cloud data stacks (AWS preferred: S3, Glue, EMR, Lambda/ECS/EKS, Step Functions; or GCP/Azure equivalents) and DBT or similar for transformations.

Built streaming systems with Kafka/Kinesis/PubSub; comfortable with event schemas and backfills.

MLOps experience with feature stores (Feast/Tecton), experiment tracking (MLflow/W&B), model registry, and model serving (SageMaker, Vertex AI, Triton, Ray Serve, or KFServing).

Production monitoring: data drift, model performance, alerting (Prometheus/Grafana, CloudWatch, Datadog).

Security basics: IAM, VPC segmentation, KMS, secrets, least-privilege, PII handling.

Clear communicator who can turn ambiguous wealth-use cases into roadmaps and ship.