Raunak Chowdhuri

Raunak Chowdhuri

About

Detail

Founder
San Francisco, California, United States

Timeline

work
Job
school
Education
folder
Project
flag
Award
auto_stories
Publication

Résumé

Jobs verified_user 0% verified
  • Reducto
    Founder
    Reducto
    Aug 2023 - Current (2 years 11 months)
  • NVIDIA
    Automous Driving Research Intern
    NVIDIA
    Jun 2021 - Aug 2021 (3 months)
    Designed a method of unsupervised representation learning to automatically retrieve relevant data points for AV active learning.
  • Oloren AI
    CTO
    Oloren AI
    Dec 2020 - Aug 2023 (2 years 9 months)
  • MIT Driverless
    Machine Learning Researcher - Perception Team
    MIT Driverless
    Oct 2020 - Sep 2021 (1 year)
  • MIT Computer Science and Artificial Intelligence Laboratory CSAIL
    Research Assistant
    MIT Computer Science and Artificial Intelligence Laboratory CSAIL
    Sep 2020 - Sep 2021 (1 year 1 month)
    Developing new methods for transferable black-box adversarial perturbations for machine learning.
  • MIT Lincoln Laboratory
    Machine Learning Research Consultant
    MIT Lincoln Laboratory
    Sep 2020 - Sep 2021 (1 year 1 month)
    Collaborating in a high-stakes multi-shareholder process to deliver aerial LIDAR landslide predictions to assist in the creation of environmental solutions for Colombia.
  • Carnegie Mellon University
    Machine Learning Research Lead
    Carnegie Mellon University
    Jun 2020 - Sep 2021 (1 year 4 months)
    - Leading a small research team to develop machine learning algorithms for music classification with deep learning, iterating on my prior independent research project, PhonoNet.
  • DeGirum
    Verification Consultant
    DeGirum
    Mar 2020 - Jun 2020 (4 months)
    - Verifying hardware modules for a machine learning accelerator startup company - Creating and testing hardware block designs using Verilog
  • P
    Machine Learning Research Consultant
    Peptide Logic
    Jan 2020 - Sep 2021 (1 year 9 months)
    - Designed a pipeline for mouse behavior classification with deep learning to automate data annotation for a neuroscience drug discovery program supported by NIH Grants. - Ongoing efforts toward a first-author publication covering the novel techniques in this approach.
  • Caltech
    Research Assistant
    Caltech
    Jun 2019 - Aug 2019 (3 months)
    - Conducted research using deep reinforcement learning for leader-follower swarm control with drones. - Periodically presented my project to the research sponsor, Raytheon. - Wrote a research proposal, shared progress at weekly meetings, and gave a final oral presentation.
  • P
    Independent Research
    PhonoNet Research Project
    Apr 2018 - Jun 2019 (1 year 3 months)
    - Conducted research work for hierarchical machine learning systems applied to Indian Classical Music and achieved highest recorded raga prediction accuracy for Hindustani Classical Music - Currently working on a system to optimize traditional epoch-based supervised learning methods using a technique inspired by spaced-repitition
  • IEEE Robotics and Automation Society
    IROS 2018 Reviewer
    IEEE Robotics and Automation Society
    Mar 2018 - Apr 2018 (2 months)
  • UNIVERSITY OF CALIFORNIA BERKELEY
    Undergraduate Researcher - Berkeley DeepDrive
    UNIVERSITY OF CALIFORNIA BERKELEY
    Apr 2017 - Sep 2017 (6 months)
    - Researched a novel approach to Multi-Task and Multi-Modal Learning through the use of privileged modal insertion allowing for distinctive individual behaviors to form as well as planning to be implemented in an end-to-end autonomous driving system. - Adapted a modular SqueezeNet network from performing the ImageNet classification task to perform a driving-like regression task with modal insertion, achieving new standards of performance. - Wrote the primary PyTorch and H5PY Training Repository in active use by the model car division at DeepDrive. This approach modularized the training repository’s data handling and allowed simple and custom training through swappable JSON config files for different training setups. - Migrated Autonomous RC
  • R
    Software Engineer - Intern
    Rockley Photonics Inc
    Jun 2016 - Aug 2016 (3 months)
    - Verified the CMOS of the Gen 1 optical network switches by creating C++ models of all the switch components. - Created efficient software for routing table generation using a modified Dijkstra’s algorithm for distributed next hop calculation in a dynamic graph.
Education verified_user 0% verified
  • Massachusetts Institute of Technology
    Bachelor of Engineering - BE, Computer Science and Electrical Engineering
    Massachusetts Institute of Technology
    Jan 2020 - Dec 2024 (5 years)
  • W
    High School, High School/Secondary Diplomas and Certificates
    Westview High School
    Jan 2016 - Dec 2020 (5 years)
    Relevant Coursework: AP Physics 1-2, AP Physics C Mechanics + E&M, AP Calculus AB + BC Westview Robotics - Mechanism Design Lead, Computer Vision Lead, Kickoff Lead USACO Computing Olympiad - Platinum (Top) Division Internationally Competitive Programming Club - Vice President Varsity Cyber Security Member - Outstanding Academic Achievement California Senate Award
Projects (professional or personal) verified_user 0% verified
  • P
    PhonoNet: Multi-Stage Deep Neural Networks for Raga Identification in Indian Classical Music
    Jun 2018 - Current (8 years 1 month)
    Research paper in progress; planned submission to ACMICMR 2019 in January. Created a hierarchical machine learning system which can handle long segments of audio data, provide live feedback which improves over time, and distinguish between more than thirty Indian classical with a state-of-the-art 99% accuracy. The system is also adaptable and can potentially extend to other audio processing tasks or other tasks with long temporal sequences as input.
  • R
    Research Blog
    I review, explain, and simplify cutting edge techniques in AI research and share progress on projects and libraries that I'm developing.
Awards verified_user 0% verified
  • A
    3rd Place Acoustical Science Award
    Acoustical Society of America
    May 2019
    $600 award at Intel International Science and Engineering Fair
  • S
    3rd Place Robotics and Intelligent Machines
    Society for Science and the Public
    May 2019
    $1000 award for novel contribution to the field of robotics and artificial intelligence at the Intel International Science and Engineering Fair.
  • GoDaddy
    Best Use of Data Award
    GoDaddy
    May 2019
    $1500 award for Best Use of Data at the International Science and Engineering Fair
  • I
    Kyoto Prize Symposium - $10,000 Scholarship
    Inamori Foundation
Publications verified_user 0% verified
  • I
    Multi-Modal Multi-Task Deep Learning for Autonomous Driving
    IEEE WACV
    Sep 2017
    A novel modal insertion method is implemented to allow a single deep neural network to learn several distinct “behavioral modes” of operation simultaneously for autonomous driving.
  • I
    PhonoViz: Chroma Feature Visualization for Hindustani Classical Music
    In Review for FMA
    A visualization technique for saliency information retrieval in convolutional networks handling audio data using LibROSA, PyTorch, and Google Compute Engine for processing.
  • A
    PhonoNet: Multi-Stage DNNs for Raga Identification in Indian Classical Music
    ACM ICMR
    A general-purpose machine learning architecture for processing long temporal sequences like those present in Indian Classical Music, achieving highest recorded accuracy in Hindustani Classical raga prediction.
This is a community-created genome.