Air‑gapped ML & AI Platform
AWS · Regulated Customer
Led design and implementation of an open‑source ML platform for highly sensitive data in a highly customized, air‑gapped environment. Responsible for architecture and strict security considerations including "secure by design" principles, threat modeling, and security testing. Established design‑forward processes that ensured delivery excellence, which directly enabled four additional double-digit-million-dollar extensions.
ADAS MLOps Platform
AWS · German Automotive
Led a team of consultants with 50% hands‑on engineering to deliver a high-scale landing zone and EKS‑based MLOps environment for computer vision pipelines (SageMaker, EMR) orchestrating massive‑scale ADAS workflows. Architected multi-tenant networking for 100Gb rapid data ingestion from copy stations, and engineered a low‑latency (<50ms) authentication mechanism. Managed full SDLC including budgeting, staffing, and automated AWS account vending.
Global EKS & CI/CD Ecosystem
AWS · Automotive
Designed an enterprise EKS platform capable of handling over 1 billion requests per day, currently hosting 160+ clusters worldwide (including China) for 700+ developers. In parallel, designed and implemented an automated software delivery platform comprising 600+ auto‑generated CI/CD pipelines to securely roll out services across multiple global locations.
AWS SSO & IPv6 Rollout
AWS · Enterprise Automotive
Designed and rolled out a serverless solution to automatically manage AWS SSO permissions, handling events like account creation, migration, OU CRUD operations, and tag‑based actions—reducing operational effort by 30% and drastically minimizing human errors. Additionally, conceptualized and prototyped an IPv6 networking cloud architecture rolled out across 3,000+ AWS accounts.
