At Tata Steel, I worked on an inventory forecasting system that used historical inventory data to forecast stock requirements.
I built AI SkillFit, a multilingual AI voice-interview platform for skill assessment, with a team. I built its face-recognition and interview anti-cheat system, and owned platform security by implementing authentication hardening and input sanitization.
I published llm-redacted on PyPI, a Python tool that detects and redacts secrets and personal information from LLM output. It supports reversible redaction, streaming sanitization, and custom regex detectors.
I also built FairAI Studio, a bias-auditing tool for binary decision models, with fairness audits and mitigation options. AI SkillFit won the AI for Bharat Hackathon.

