Location: San Francisco, CA
Remote: [Yes / Hybrid / On-site SF]
Willing to relocate: Yes for NYC and London
Technologies: Python, C++, SQL, AWS (CDK, CloudFormation, Lambda, EventBridge, DynamoDB), Docker, Kubernetes, CUDA, PyTorch
Résumé/CV: email me on the address below.
Email: [email protected]
Infrastructure and backend engineer, 3.5+ years. I work on the Amazon Redshift control plane at AWS.
I built the production control-plane backend for Redshift Serverless's AI-driven autoscaler: resize orchestration, failover, and state management. I'm the person who makes ML systems run reliably in production, not the one training the models.
Other work at AWS:
- Cut deployment blast radius ~95% with fault isolation on a core orchestration service. Zero rollbacks.
- Cut multi-region deployment time from 3-4 weeks to 5-7 days by moving core services to AWS CDK.
- Led the Graviton4 rollout for Redshift Serverless. ~30% per-node compute gains, zero customer impact.
Before that I shipped a CUDA MRI reconstruction solver and a 3D U-Net on Vertex AI at a medical imaging startup. MS EE, Columbia.
Looking for: AI/ML infrastructure, platform, or backend roles at Series A/B AI companies.
I built the production control-plane backend for Redshift Serverless's AI-driven autoscaler: resize orchestration, failover, and state management. I'm the person who makes ML systems run reliably in production, not the one training the models.
Other work at AWS: - Cut deployment blast radius ~95% with fault isolation on a core orchestration service. Zero rollbacks. - Cut multi-region deployment time from 3-4 weeks to 5-7 days by moving core services to AWS CDK. - Led the Graviton4 rollout for Redshift Serverless. ~30% per-node compute gains, zero customer impact.
Before that I shipped a CUDA MRI reconstruction solver and a 3D U-Net on Vertex AI at a medical imaging startup. MS EE, Columbia.
Looking for: AI/ML infrastructure, platform, or backend roles at Series A/B AI companies.