Engineering manager for the layer between research and customers: models into products, production signal back into training.
- Agent experience: the model factory’s agentic layer, rolling research improvements out across the organization
- Managed agents: the enterprise Console, with RBAC, policies and full session traceability
- Evals and self-improvement: customer-specific eval environments, PostTrainBench and synthetic RL environment generation
- Fine-tuned models: shipped an end-to-end fine-tuning stack for customer-specific training
- Inference and cloud: Atlas models on AWS Bedrock, and Poolside’s cloud offering
Go · TypeScript · Python · PyTorch · vLLM · Kubernetes · Postgres · Dagster