Roger Scott

I’m a machine learning and platform engineering leader who builds production ML systems, data platforms, and AI-driven applications. I focus on scalable model serving, monitoring, and data pipelines that work reliably in enterprise environments.

At Anaconda, I directed development of the Anaconda AI platform, managing 9 engineers and delivering Kubernetes-based AI services and inference systems. I led an LLM inference platform with model quantization, scalable API endpoints, and RAG-based systems using PostgreSQL embeddings.

Previously at Mastercard, I led 3 engineering teams and built a machine learning model serving platform with orchestration, A/B testing, multi-arm bandit testing, and SHAP-based explainability. I also designed and deployed an ML inference API that improved payment authorization rates by 15–20%, while implementing real-time model monitoring for drift and reducing deployment time by ~40%.

Earlier roles strengthened my end-to-end instincts across data engineering and platform architecture. At Jane.com, I led production pipelines using AWS Glue and Athena, migrated event tracking to Segment, and delivered recommendation and attribution systems tied to measurable revenue and efficiency gains; as a Senior Architect/Technical Director, I modernized platforms with Java Spring Boot microservices and improved reliability by migrating Kafka to AWS SQS.