Sharifi Renani

Senior Machine Learning Scientist at Optum

Generative AILarge Language Models (LLMs)Foundation ModelsAgentic SystemsBiomechanicsTime-Series Analysis

About

I'm Sharifi Renani, currently a Senior Machine Learning Scientist at Optum where I focus on Foundation Models and Agentic Systems. My career has been a journey of bridging the gap between physical engineering and advanced artificial intelligence. With a Ph.D. centered on deep learning for biomechanics and experience at companies like Spotify and Zimmer Biomet, I’ve worked on everything from recommendation systems and generative AI to orthopedic implant design. I am deeply passionate about open science and using synthetic data to solve the data bottlenecks in specialized fields like healthcare. Whether it's developing the BioMAT model for gait analysis or exploring music-understanding LLMs, I thrive on applying complex mathematics to real-world problems. I’m always looking to connect with researchers and professionals in the AI and digital health space to exchange knowledge on clinical data and innovation.

Networking

What I can offer

›Expertise in bridging complex physical systems with advanced AI.
›Deep knowledge of LLMs and recommendation systems.
›Insights into healthcare AI and wearable sensor analysis.

Looking for

›Best practices for clinical data collection
›Collaborations on hospital-based surveys
›expanding my professional network
›exploring mutual opportunities in Healthcare AI and Machine Learning

Best fit for

AI research communityHealthcare professionalsMusic-tech industry leadersStartup founders and VC investors

Current Interests

Generative AI for AudioDigital Health and WearablesComplexity and Physics in Computational ModelingVenture Capital and Startup EcosystemsValue-based Healthcare Models

Background

Career

Transitioned from a background in Mechanical Engineering and Biomechanics into deep learning research, moving from academic research and internships at Zimmer Biomet and Spotify to senior ML roles focusing on Foundation Models and Agentic Systems.

Education

Ph.D., University of Denver (2018 – 2021); M.Sc. in Biomechanics, University of Missouri-Kansas City (2015 – 2017); M.Sc. in Mechanical Engineering, The University of Toledo (2014 – 2015); B.Sc. in Mechanical Engineering, Isfahan University of Technology (2009 – 2014).

Achievements

›Developed and published the open-source BioMAT Model for joint kinematic predictions.
›Developed a methodology using GANs and VAEs to generate synthetic kinematic data.
›Achieved 90% accuracy in spatial-temporal gait prediction.
›Authored two books on computational models for human organs.
›Led design reviews for the Vanguard® 360 Revision Knee System.

Opinions

Open-source datasets and models are essential to advance scientific research.
Synthetic data is a vital solution to data acquisition bottlenecks in specialized domains.
Burnout among orthopedic residents is a critical issue that needs addressing.