Hey, I'm Ankush.

I work on building AI-assisted and backend systems within AWS, focusing on developer productivity, cost optimization, and platform reliability. My work includes automating on-call handoff documentation using GenAI, optimizing canary executions to reduce infrastructure costs, and delivering end-to-end integration tests for key features in the SageMaker Unified Studio (SMUS) portal. I also contribute to AI-assisted data classification workflows in SageMaker Catalog, improving metadata quality and governance at scale.

I develop and implement AI-driven solutions, including advanced document chunking strategies using LlamaParse and OpenAI GPT-4. I created a Retrieval-Augmented Generation (RAG) evaluation set utilizing Anthropics' Contextual Retrieval methods, tested various embedding models for performance optimization, and built APIs with FastAPI for seamless integrations.

I help build and enhance the Ascella Groups Page. By implementing verticals and animations, I significantly improved the user experience. Additionally, I optimized the frontend performance, resulting in a smoother and more responsive interface. My work ensures that users enjoy an engaging and efficient experience on the platform.

I analyzed booking patterns and optimized service availability, leading to a 12% reduction in customer wait times. By implementing process improvements and automation solutions, I achieved a 15% increase in productivity. I collaborated with cross-functional teams to gather requirements, define project scopes, and ensure alignment with business objectives, fostering effective teamwork and project success.