Mihir Gujarathi — Machine Learning Engineer

I’m a Machine Learning Engineer with a Master’s in Machine Learning and Data Science from UC San Diego. My passion lies in building AI systems that combine research and innovation with a focus on real-world impact. At UC San Diego Health, I worked on AI solutions tailored for clinicians. I built a RAG-based chatbot that supported single-cell analysis using the ScanPy library, developed a speech transcription pipeline with Whisper that reduced ~30 hours of manual transcription, and designed a lightweight CNN for detecting ventricular arrhythmias on edge devices, achieving 33 ms latency with an 80% smaller model. In Summer 2024, I interned at Causify, where I designed a data streaming pipeline with Airflow, Spark, and Cassandra for their KaizenFlow framework, and improved its robustness by increasing test coverage by 10%. Earlier, at the Rajiv Gandhi Science & Technology Commission, I developed a hybrid CNN+RNN architecture for Indian Sign Language recognition, improving accuracy by 13% through spatiotemporal features and human body keypoints. I’m currently exploring LLM inference optimization, Agentic AI, and on-device machine learning — areas that challenge me to push the limits of efficiency while still delivering meaningful results. Outside of engineering, you’ll likely find me hiking, surfing, or justifying debugging as a legitimate hobby. Let’s connect if you’re looking for someone who enjoys solving tough problems and turning AI ideas into production-ready solutions.

Stackforce AI infers this person is a Machine Learning Engineer with a focus on Healthcare and Data Engineering.

Location: San Diego, California, United States

Experience: 9 mos