DataRobot offers an end to end AI platform for data scientists and engineers of all skill levels to build and deploy predictive, generative and agentic AI workflows. I've been part of and lead the MLOps and Governance teams allowing users to operate and scale their AI workflows in any environment with bolt-on governance features allowing users full insights and confidence into their workflows. My biggest contribution was being the lead developer and architect behind DataRobot's batch prediction functionality allowing users to integrate the ML models into their own data/ETL pipelines at any scale. I then stepped back into a management role, leading teams through adoption of a serverless architecture for a cost-effective and reliable way of serving AI workloads as well as re-architecting our monitoring/governance capabilities onto the industry standard OpenTelemetry protocol — allowing for seamless integrations.
While my role has transitioned into a Director role, I never stepped fully away from coding and debugging. I try to find time to still have an impact through smaller improvements as well taking an active role in our Incident Response program where I'm a frequent participant as a responder. I find incidents to be huge learning opportunities with short feedback loops — and I'm always curious how we can improve.
Technologies: Linux, Kubernetes, Nginx, Python, Go, OpenTelemetry, Prometheus, Pandas, Numpy, uWSGI, Flask, FastAPI, grpc, protocol buffers, RabbitMQ, MongoDB, PostgreSQL, Snowflake, Synapse, BigQuery, AWS, Azure, GCP, git