Sushil Kumar

• Architected a secure, hybrid Generative AI platform leveraging Azure OpenAI GPT-4 isolated in a dedicated private Azure tenant, successfully integrating legacy systems to achieve a 75% reduction in operative note documentation time. • Introduced ChatGPT-assisted coding and GitHub CoPilot to accelerate development, reducing release cycles and elevating code quality. • Mentored teams on LLM prompt engineering, agent-based automation, and GenAI design patterns, embedding reusable architectural components.

• Spearheaded LLM-driven workflow pilots and initiated early RAG pipeline implementations for knowledge retrieval within a complex financial corpus, establishing architectural foresight for advanced semantic understanding. • Managed MLOps governance and delivery across a 35-member engineering team, establishing scalable AWS AI/ML infrastructure foundations (SageMaker, S3) and enforcing CI/CD process consistency via SageMaker Pipelines and the Model Registry. • Coaching engineering leads on latest trends.

• Founded and led a smart-farming venture leveraging IoT-enabled monitoring and predictive analytics. • Designed cloud-connected infrastructure to optimize crop yield and resource efficiency. • Launched web and mobile applications in collaboration with distributed product teams, integrating supply-chain automation and data visibility.

• Directed global modernization programs implementing AWS migrations, ML‑assisted analytics, and automation. • Architected cloud-native ML environments supporting data pipelines, ETL modernization, and analytics workloads. • Introduced DevOps and MLOps foundations: CI/CD, IaC via Terraform, containerization, automated QA, and monitoring. • Notable customers: National Grid, Lloyds Banking Group, State Street, Corning, Xerox, Stryker etc.

• Designed GPS/GPRS-based tracking systems with data encryption and tiered access roles in fleet and tolling domains. • Re-architected high-throughput modules in Java; gained hands-on exposure to systems that later influenced my security thinking.