Founding Sanofi’s first AI engineering organization
Founded Sanofi’s first AI engineering organization and scaled it to 70+ people across Toronto, Paris, and Boston — 7 direct reports, 30+ FTEs, 40+ contractors — reporting to the VP of R&D.
- Defined and executed the global AI infrastructure strategy: a $12M+ HPC platform on NVIDIA A100/H100, EKS, FSx Lustre, and S3 — securing dedicated infrastructure over a competing corporate platform through alignment with R&D, IT Security, Legal, and Finance.
- Delivered platform architecture and budget strategy to SVP stakeholders within 6 weeks. Blueprints and Terraform modules adopted by Vaccines and Small Molecule R&D as Sanofi’s internal reference.
- Shipped an event-driven LLM platform for automated clinical data analysis: vLLM-served models with FP8 quantization on ECS, EventBridge, and AWS Batch — manifest-driven orchestration that scales to zero between workloads, eliminating idle GPU spend.
- Standardized Terraform IaC and automated provisioning pipelines: 23% throughput improvement and reproducible ML experimentation at scale.
- Operationalized protein models via SageMaker and Bedrock, reducing time-to-production 26% across 7+ R&D teams.
- Cut incident frequency 50%+ and onboarding time 40% through standardized CI/CD and async-first documentation.
Stack · AWS · EKS · ECS · EventBridge · AWS Batch · FSx Lustre · S3 · vLLM · SageMaker · Bedrock · Terraform · HPC (A100/H100)