FUll Time
FUll Time
Embedded Files
Part of Audio Algorithms team in Media Foundation Core org
Developing Audio AI models for apps like Messenger and Whatsapp, serving billions of users worldwide
Sr. Audio AI Engineer at QSC
Sr. Audio AI Engineer at QSC
Date: May 2023 - November 2025
Mentors: Matt Skogmo
Developed a novel Multichannel Source Separation model based on Conformer architecture that outperforms available models especially in far-field scenarios involving multiple talkers and noise sources in meeting room environments
Compressed the model using Pruning, Quantization, Knowledge-Distillation, Multitask learning and spearheaded deployment of separation model on edge device, demonstrating real-time streaming capabilities and laying the foundation for future models
Designed and implemented versatile Dataset creation pipeline from scratch, having functionalities like pre-processing, augmentation and mic emulation which enables creation of synthetic datasets for range of Audio AI tasks including Source Separation, Noise classification, Direction of Arrival Estimation, and Dereverberation
Delivered impactful presentations and demos to senior leadership and colleagues, increasing visibility and momentum of Audio AI projects within the company, prepared demos for multiple sales events
Published two patents
Internships
Internships
Audio Machine learning coop at Bose
Audio Machine learning coop at Bose
Date: July 2022 - Present
Mentors: Alex Pawlicki
Worked on the following two major projects:
Three-STEM Audio Source Separation: Coded online input mixing pipeline for audio data of more than 200 TB and trained Multi-resolution cross net model from cocktail fork paper on it. Engineered model to be causal without any performance loss in SI-SNR values by using LSTM instead of Bi-LSTM layers and adding Self Attention layers.
Gender Bias Free Speech Enhancement (SE): Developed a new version of Bose’s dataset by adding gender labels predicted using 95% accurate novel Deep Learning model trained on BYOL-A audio embeddings. Made significant changes in existing codebase to develop training scheme which improved performance of SE model for under-represented gender class by 3% without loss in performance of other class.
Computer Vision Intern at Spyne.ai
Computer Vision Intern at Spyne.ai
Date: January - May 2021
Mentor: Mr. Deepanshu Singh
Designed a pipeline with 40% increase in the performance over previous methodology, for automated bulk processing of car images for websites like OLX, Cars24, etc. (Video link, Credits: Spyne.ai team)
Replaced noisy background in the car images with Showroom background using AI model
Developed a novel U-Net based Generative Adversarial Network for realistic shadow and reflection generationin car images with replaced background
Detected license plate regions in the car images and replaced them with company logo
Added pre- and post-processing steps using Digital Image Processing techniques to enhance results
Coordinated with software development team to generate APIs and test them using Postman platform
research Intern at AI Institute, University of South carolina
research Intern at AI Institute, University of South carolina
Date: March - August 2020 (Third year of B.Tech)
Mentors: Prof. Amit Sheth, Dr. Shweta Yadav
Worked on the project of Depression severity detection from Tweets
Implemented Multilabel classification task using state of the art language models like BERT as baseline models
Developed novel FiLaMTL - figurative language enabled multitask learning framework to achieve improvement of 4% above state of the art models
Long research paper published at COLING 2020 conference (rank-6, h5-index 49)
Winter Internship at PHotonics lab, IIT Gandhinagar
Winter Internship at PHotonics lab, IIT Gandhinagar
Date: December 2020 (Third year of B.Tech)
Mentors: Prof. Aruplal Chakraborty, Prof. Nithin V George, Chandankumar Jha
Developed a mic using FBG sensor, which can reduce the background noise by significant amount
Performed several experiments to get frequency response, PESQ score of the recorded signal and for the performance enhancement of the FBG based mic
A chieved 30% better signal quality than conventional mic in noisy environment
Additionally applied several signal processing techniques like Adaptive filtering, High Pass Filtering and Singular Value Decomposition over the speech signal recorded by FBG based mic to make it more intelligible
Gained hands on experience of sensor development and signal processing while working on this project
Research paper accepted at 27th Optical Fibre Sensors - OFS conference
Also volunteered for PHASE 2020 workshop organised by the Photonics lab IIT Gandhinagar in colaboration University of Strathlyde and some industries
Summer Intern at Addverb Technologies
Summer Intern at Addverb Technologies
Date: May - August 2019 (Second year of B.Tech)
Mentors: Mr. Manish Jha, Mr. Vikash Kumar
Implemented path planning and path tracking for mobile robots
Studied and simulated the “Pure pursuit” algorithm for autonomous vehicle’s path tracking to differential driven sorting robot in Gazebo-ROS using nodes written in Python
Integrated a code for robot’s smooth velocity profile to reduce the jerk in the above implementation
Created a template code for Heat map, based on robot’s position to generate weights for “A-star” algorithm-based path planning to avoid collision between 300 robots working together
Mastered ROS framework, Networkx module for Python, object-oriented programming and threading in Python and simulation software like Rviz and Gazebo
Gained exposure to industry 4.0 automation and projects of many multinational companies and startups in this field while attending Indian Warehouse Show-IWS 2019 as a member of Addverb Technologies
Summer Intern at ONGC
Summer Intern at ONGC
Date: May - August 2018 (First year of B.Tech)
Mentor: Mr. Krishnan
Studied electrical machines and transmission equipment used at Hazira plant of ONGC
Prepared a report on power requirements of plant, power generated at the plant using Gas turbine generators, power supplied by the plant to the grid
Also studied stepping up and down the voltages at substations and their control
Acquired exposure to industries so that I can work on my research projects more accurately and they can be useful to the society
Travel Fellowship
Travel Fellowship
Availed explorers fellowship offered by IIT Gandhinagar
Availed explorers fellowship offered by IIT Gandhinagar
Date: May - August 2018 (First year of B.Tech)
Spent 47 days, travelling continuously in the budget of 37,800Rs.(532$) and visited around 38 places all over India on the theme "unexplored India"
Explored 10 unique unexplored places of India inluding Hampi, Majuli, Rumtek etc. and wrote a blog on it to make people aware about this treasure of India
Won 2nd prize in the same fellowship for the blog writing
Positions of responsibilities
Positions of responsibilities
Teaching Assistant, Faculty Development Program in Machine Learning
Teaching Assistant, Faculty Development Program in Machine Learning
Date: January 2021
Helped more than 45 faculties from all around India in learning the basics of Python and Data science
Created Jupyter notebooks demonstrating basic tutorials of Python and its different libraries
Teaching Assistant, NLP course, IIT Gandhinagar
Teaching Assistant, NLP course, IIT Gandhinagar
Date: August - December 2021
Initiated two NLP projects "Bhagavad Gita based chatbot for ethical dilemma question-answering" and "AI framework to generate jokes along with GIFs", and mentoring two groups on these projects
Framed questions, graded them and provided Prof. Mayank Singh logistic support needed
Academic discussion hours mentor, electronic devices course, iit Gandhinagar
Academic discussion hours mentor, electronic devices course, iit Gandhinagar
Date: August - December 2021
Guided students on their queries related to Electronic Devices course as a part of Earn While You Learn initiative of IITGN to promote Peer Assisted Learning
Secretary, Media and PR cell, iit Gandhinagar
Secretary, Media and PR cell, iit Gandhinagar
Date: May 2019 - August 2020
Leading a team of 10 members including content writers and photographers
Increased the reach of the IIT Gandhinagar Student life Facebook page(with 5000+ followers) by 20% by posting regularly about the activities going on in the campus and making people aware of it
Coordinated with Chief Editor of newspaper “Pal Pal”, Mr. Surendra Bhatia to cover salient features of the campus in media platforms like newspapers, as well as YouTube and news blogs
Organised two events, Roboquest and Drift in Amalthea ’18 successfully which were also covered in reputed newspapers like Times of India, which built leadership quality and spirit of teamwork in me
Coordinated with other 80 team members from different departments to successfully organize the event
Organised soft-skills workshop for the student community
Compiled and systemized interview preparation material
Leading the focused group of robotics at IIT Gandhinagar to bring together like-minded people and provide them necessary resources
Coordinating with representatives of other prestigious colleges of Ahmedabad and Gandhinagar like IIM Ahmedabad, DAIICT, PDPU, NID etc. for various events and activities
Acting as a point of contact between IIT Gandhinagar and other colleges to provide the student community with relevant and helpful information from other colleges
Page updated
Google Sites
Report abuse