Shubham Khavate

I'm currently helping GlideCloud modernize legacy mainframe applications through AI-assisted code transformation, code analysis, and automation workflows.

I work on converting existing mainframe business logic into modern, maintainable software architectures while improving migration, documentation, validation, and code-understanding processes.

I built BrandToneChecker, an NLP application that evaluates content against predefined brand tones using vector embeddings, semantic similarity search, ChromaDB, FastAPI, and Llama-powered rewriting. I also developed CancerSpy, a CNN-based skin cancer detection system that achieved 99.7% training accuracy and 98.7% testing accuracy.

Alongside my engineering work, I led 20+ members as Event Management Head for the Artificial Intelligence Students’ Association and organized Vizathon, a data-visualization hackathon with 200+ participants.