Experience
My professional journey spanning research, computer vision engineering, and data science.
Graduate Research Assistant
Visual Intelligence Laboratory
Working in the Visual Intelligence Laboratory under Dr. Stephen Baek, conducting research at the intersection of computer vision and biomechanics. Primary focus on a NIOSH-funded grant evaluating markerless motion capture for occupational ergonomics, alongside research projects in biomechanical angle estimation and sports performance analysis. Also serving as a teaching assistant across multiple graduate-level data science courses.
Key Responsibilities
- Leading research on a NIOSH Grant evaluating vision-based approaches for worker posture analysis, comparing stereo and monocular camera systems against gold-standard motion capture across a 48-subject study with simulated manufacturing tasks
- Developing methods to convert 3D pose estimation outputs into biomechanically meaningful joint angles following ISB standards, bridging the gap between computer vision and clinical biomechanics
- Building an automated pipeline for analyzing tennis serves from internet video, including serve detection, 3D pose extraction, and biomechanical metric computation at scale
- Improving 2D-to-3D lifting capabilities of human pose estimation algorithms using geometric and physical constraints, and leveraging generative AI for synthetic pose data augmentation
- Teaching assistant for graduate courses including Bayesian Machine Learning, Deep Learning, Numerical Analysis and Optimization, and Probability & Stochastic Processes
Technologies & Skills
Key Outcomes
- Published and submitted research on vision-based posture analysis methods to IISE Transactions
- Developed novel pipeline converting pose estimation skeletons to ISB-standard biomechanical angles
- Created large-scale dataset of thousands of annotated professional tennis serves with automated 3D analysis
- Presented at international conferences including ISB (Stockholm) and ACSS (Singapore), winning multiple presentation awards
- Taught across 4 graduate data science courses over 4 semesters
Computer Vision Engineer Intern
Biocore LLC
Contributed to a machine learning system for 3D human pose reconstruction of NFL football players, working with multi-camera setups and deep learning models to track player movements. Collaborated with the NFL's AWS Next Gen Stats infrastructure to support on-field safety analysis.
Key Responsibilities
- Performed camera calibration across 30+ stadium views to enable accurate multi-view triangulation and 3D pose reconstruction of football players
- Trained and fine-tuned deep learning models to improve pose estimation accuracy under heavy occlusion scenarios common in game footage
- Contributed to the end-to-end ML pipeline integrating player detection, pose estimation, and 3D reconstruction into a unified system
- Collaborated with NFL's AWS Next Gen Stats system to analyze player biomechanics and enhance on-field safety metrics
Technologies & Skills
Key Outcomes
- Successfully reconstructed 3D player poses from 30+ simultaneous camera views
- Improved pose estimation accuracy in heavily occluded game scenarios through model fine-tuning
- Contributed to NFL player safety and biomechanics analytics pipeline
Data Science Intern
Gamebytes (YC W19)
Drove data-driven growth strategy at a Y Combinator (W19) backed iPhone social messaging startup, leveraging user behavior analytics to identify key retention drivers and inform product decisions during a period of rapid scaling.
Key Responsibilities
- Designed and executed user retention analyses to uncover behavioral patterns correlated with long-term engagement, directly informing product strategy
- Built data pipelines using Python and SQL to extract, clean, and analyze large-scale user activity databases, transforming raw data into actionable insights
- Developed predictive models to segment users by engagement likelihood, enabling targeted feature development and personalized onboarding flows
Technologies & Skills
Key Outcomes
- Contributed to growth strategy that propelled the app from #120 to #3 in the App Store Social Media category
- Identified key user activity signals predictive of long-term retention
- Delivered actionable recommendations that shaped product roadmap and feature prioritization