Machine Learning DevOps Engineer

About us:
Boerboel Trading is an electronic trading firm, active across a wide range of asset classes and regions, with offices in NYC, Chicago, London, and Malta. We engage in rigorous quantitative analysis to deploy systematic trading strategies over multiple time horizons. Trading and research employees of the firm are graduates from a wide spectrum of quantitative disciplines, such as engineering, computer science, mathematics and physics. Our employees apply their scientific methodologies to the analysis of extensive financial data sets.

Summary:
We operate high-performance infrastructure processing 20PB+ of market data. We're building a production ML platform to organise features, orchestrate model training, and serve predictions for systematic strategies. As a ML DevOps Engineer you'll design and implement the feature store, data lake organisation, and ML pipelines that enable our research and trading teams to work efficiently at scale.

Objectives:

• Feature store & data lake: Build scalable infrastructure for time-series feature storage, retrieval, and versioning optimized for ML workloads
• MLOps pipelines: Design end-to-end workflows for data ingestion, feature engineering, model training, backtesting, and deployment
• Data ingestion layer: Connect raw data streams into structured, queryable formats (Parquet/Delta Lake)
• Production serving: Deploy feature computation and model inference with appropriate latency characteristics
• Integration: Work with existing data capture and execution systems

Experience:

• Python depth: Strong Python engineering with focus on data pipelines and ML infrastructure
• ML platform experience: Built feature stores, MLOps pipelines, or similar systems at scale (Netflix, Yandex, Criteo, or comparable tech companies)
• Data lake expertise: Experience ingesting, organising, and serving large datasets (TB-PB scale)
• Production mindset: Deployed systems with reliability, monitoring, and performance requirements
• Distributed computing: Familiarity with Spark, Dask, Ray, or similar frameworks

Required skillset:

• Python-first with some C++ integration points
• Modern data stack: Parquet, Data Lake, distributed compute
• Your choice of orchestration tools (Airflow, Prefect, etc.)
• 20PB storage, nanosecond-resolution data, real-time streams

Preferred:

• C++ exposure (helpful for integration but not required)
• Real-time or time-series data experience
• Knowledge of Data Lake, Iceberg, or modern table formats