Summary
Tarpan Mishra is a machine learning engineer with a decade of experience building cloud-native, production-grade AI systems and backend services in the San Francisco Bay Area. He designs and deploys scalable microservice architectures and inference pipelines—leveraging FastAPI, Celery, Kubernetes, and cloud providers—to support high-throughput video analysis with sub-300 ms frame processing. His work blends applied ML (XGBoost, random forests) with real-time data engineering (Kafka, Redis, PostgreSQL) and domain-specific solutions from biomechanics to water infrastructure monitoring. At FormIQ he introduced retrieval-augmented agents using LangChain for adaptive exercise insights, and his research background includes using simulation and deep learning to detect cyber attacks on water systems. A UC Berkeley applied math and CS student by training, he pairs rigorous quantitative foundations with hands-on deployment skills and an active npm presence (tarepan) that signals ongoing open-source experimentation.
9 years of coding experience
Applied Mathematics and Computer Science, Applied Mathematics and Computer Science at University of California, Berkeley
High School Diploma, Computer Science, High School Diploma, Computer Science at Dougherty Valley High School
English, Odia