cv
General Information
| Full Name | Prarthana Bhattacharyya |
| Languages | English, Bengali, Hindi, Japanese |
| Work Authorization | UK Global Talent |
Experience
-
2025 - Present Machine Learning Engineer
Eedi, London, UK - Developing and deploying ML and Gen-AI solutions for education technology across UK schools.
- Co-authored research on uncertainty-based approach for identifying struggling students.
- Refactored Knowledge Tracing (KT) pipeline, reducing code by 31% and implementation time by 80%.
- Deployed a generative next-question selector for dynamic quizzes in pilot schools. Productionized on Azure using LitServe, Terraform, and GitHub Actions.
- Led internal workshop on Generative AI for Education and LearnLM, building team capability in LLM applications.
-
2023 - 2025 Senior Machine Learning Engineer
Ultraleap, Bristol, UK - Led hand tracking foundation model development for AI music tutors, redesigning core architecture within 4 months for improved temporal modeling and 3D spatial accuracy.
- Developed ultra-low-power gesture control for smart glasses achieving 20% accuracy improvement (F1>80%) with 25x power reduction (CES 2025, public launch).
- Built micro-gesture detection system for smart eyewear, demonstrated at AWE USA 2024 and featured by CNET (Best Paper, ECCVW 2024).
- Improved IR pinch-tracking accuracy by 50%+ through dataset curation and addressing pose-class imbalances.
-
2021 - 2022 Machine Learning Researcher
Gatik, Toronto, Canada - Designed and implemented temporal motion forecasting solutions for autonomous driving (Mitacs PhD Fellowship).
-
Summer 2022 Machine Learning Engineer Intern
Apple, Toronto, Canada - Implemented self-supervised learning for Vision Transformers, improving Visual Lookup classification accuracy by 4-6%.
-
Summer 2021 Applied Scientist II Intern
Amazon Science, Vancouver, Canada - Developed self-supervised transformer for logo recognition, achieving 4-11% improvement over SOTA (ICASSP 2022).
Skills
| Machine Learning | PyTorch, Keras, TensorFlow, Transformers, Vision Transformers, Self-Supervised Learning, Knowledge Tracing, Computer Vision, Gesture Recognition |
| MLOps & Cloud | Azure, LitServe, Terraform, GitHub Actions, CI/CD |
| Domains | Generative AI, Edge ML, Autonomous Driving, Educational Technology, 3D Computer Vision |
| Programming | Python, MATLAB, SQL, Git |
Education
-
PhD in Electrical and Computer Engineering
University of Waterloo, Canada - Developed ML algorithms for 3D object detection, tracking and motion forecasting for autonomous vehicles. 3rd place at NeurIPS 2019 Argoverse competition.
- Research deployed on autonomous vehicles operating on public roads in Canada.
- Published at top-tier conferences and workshops (ECCV, ICCV, CoRL, IV) with 450+ citations and 400+ GitHub stars.
- Mentored five undergraduate and MS students on thesis projects in 3D computer vision, self-supervised learning, and temporal motion forecasting.
-
Master of Science and Technology
University of Tokyo, Japan - Specialized in 3D scene understanding and sensor fusion for autonomous vehicles. Developed ML approach for traffic analysis at urban intersections validated with real-world Tokyo data (IEEE VTC). MEXT Scholarship recipient.
-
Bachelor of Engineering
Jadavpur University, India - Focused on digital image compression using machine learning and fuzzy clustering. Published research on vector quantization-based image compression at IEEE conference.
Honors and Awards
-
- CES 2025 - Public demonstration of hand gesture recognition system on smartglasses
- Best Paper Award, ECCV 2024 Workshop on Smart Eyewear
- AWE USA 2024 - Public demonstration of micro-gesture detection (featured by CNET)
- Mitacs Accelerate Fellowship, Gatik AI (2021-2022)
- 3rd Place, NeurIPS 2019 Argoverse 3D Tracking Competition
- Graduate Fellowships, University of Waterloo
- MEXT Scholarship by the Japanese Government, for M.S. studies in Japan