Yash Mohadikar

I’m a final-year Computer Engineering student specializing in Machine Learning and AI systems, focused on designing, deploying, and monitoring end-to-end ML pipelines. I build low-latency inference APIs and manage production ML systems with strong attention to experiment tracking and reliability.

In my AI Engineering projects, I engineered a credit card fraud detection platform that achieved 98% recall at 90% precision on a 10,000+ transaction holdout dataset. I deployed a production FastAPI inference service with < 10 ms latency per transaction and implemented 100% prediction audit logging using PostgreSQL.

I also implemented monitoring that detects model drift within 2 hours using schema validation and performance tracking. To keep deployments consistent, I built containerized ML infrastructure with Docker (including docker-compose) for dev–staging–production workflows, and tracked 10+ MLflow experiments with reproducible training pipelines.

Earlier, as a Data Science Intern, I developed 6+ AI chatbot systems, including a medical Q&A assistant trained on 47k+ MedQuAD pairs and a domain-expert bot trained on 1M+ arXiv research abstracts. I bring that same end-to-end mindset to multimodal work too—building an AI debate coach pipeline spanning STT, NLP analytics, voice pitch detection, and posture estimation.