PRITHVIJIT CHAKRABARTY

Experience

• Amazon A9 Search: Rufus AI Shopping Assistant
  (Jan 2023 - current)
  • Large scale distributed training, training LLMs from scratch
  • Optimizations for low-latency, high-throughput LLM inference
  • Controllable text generation
  • Official release video and official release announcement
  


• Amazon A9 Search
  (Jun 2021 - Dec 2022)
  • Large scale multilingual, multimodal model training
  


• Amazon Go: Dash Cart
  (Jul 2019 - Jun 2021)
  • Algorithms team for the Dash Cart project
  • On-device machine learning: computer vision, sensor processing, indoor navigation
  • Official product page and product review.
  


• Master's in Computer Science at University of Massachusetts, Amherst
  (2017 - 2019)
  • Internship: Interaction and Analytics Lab at Xerox Parc
  • (Summer, 2018) Techniques few-shot learning, training object detectors with limited labeled data
  • Mentor: Bob Price
  


• Bachelor's Degree in Computer Science at VIT University
  (2013 - 2017)
  • Internship: Intel, Bangalore
  • (2017, 6 months) Real-time classification and segmention of hyperspectral images
  • (Summer, 2016) Drive quality and event analysis with accelerometer, gyroscopic data
  • Mentors: Anbumani Subramanian, Subramanian Chidambaram
  • Internship: Stride AI
  • (Summer, 2015): Anaphora, coreference resolution.

Papers

• The Spectral Bias of the Deep Image Prior
    Prithvijit Chakrabarty , Subhransu Maji  
    Bayesian Deep Learning Workshop, NeurIPS 2019     Code     Poster



• Automatic adaptation of object detectors to new domains using self-training
    Aruni RoyChowdhury, Prithvijit Chakrabarty, Ashish Singh, SouYoung Jin, Huaizu Jiang, Liangliang Cao, Erik Learned-Miller  
    CVPR 2019     Project page     Code



• On Optimizing Human-Machine Task Assignments
    Work done as a part of the Aspiring Researcher Program (2016), with Prof. James Davis at UC Santa Cruz and Rajan Vaish
  

Reading List

I like reading books

• 2026, 2025
• 2024, 2023, 2022, 2021, 2020

Projects (old)

  GitHub

Course projects

• Selectively Forgetting Tasks with Elastic Weight Consolidation for Continual Learning   Report   Code
• Thompson Sampling for Contextual Bandits   Report

Personal
• Simulations on attractor networks for memory   Code
• Exercise: Annealed Importance Sampling   Code   Video
• Hand tracking with autoencoders   Demo   Code
• Graph centrality to find main movie characters   Demo   Code   Blog post
• CNN for face detection   Demo   Code
• Human tracking with SVMs   Demo   Code
• Sentiment analysis with Convolutional Neural Networks   Code and documentation
• 3D mapping with ultrasonic sensors   Video   Code   Documentation
• Beeble: A collaborative Android drawing app   Video   Code
• A visual music player in Processing   Code
• Po: Simple MapReduce in C   Code
• Graph plotter   Code
• Flashcards   Code

  GitHub gists



Random Fun Stuff (very old)

• PC build
• Music: I formally studied music theory and play the violin. I also play the guitar, piano and the harmonica.
• The big bad bottle wheel!   Video
• Thoughts on TempleOS   Video
• Very old snippet on graphs
• Emotional article I wrote in school