Feb 2023 – Aug 2023 · 6 mos · Pittsburgh, Pennsylvania, United States · On-site

• Meta AI sponsored research on vision and multimodal problems for AR/VR glasses.

May 2022 – Aug 2022 · 3 mos · Greater Seattle Area

• Analyzed payment failure data and improved web page notification system for payment failures to be more descriptive and provide rectification steps to customers. Reduced customer tickets regarding payment failures by 72%.

Feb 2019 – Apr 2021 · 2 yrs 2 mos · Delhi, India

• 1. I was involved in homography estimation for alignment and information extraction from document images. The work has been patented, and published in ICDAR 2019 Workshop.
• 2. I worked on medical image anaylsis problems like exploring meta-learning for classification and segmentation in medical image datasets like skin lesions and thoracic diseases from chest x-rays. Used x-ray region localization to specifically identify diseases and abnormalities. Work was in collaboration with ILBS Delhi. The work on meta-learning led to a publication in CVPR Workshop 2020.
• 3. Explored novel ways of improving Covid-19 detection from chest X-rays and other diagnostic images using minimum data. The created pipeline improved the state-of-the-art covid detection performance and was published in MICCAI 2020 Workshop.
• 4. Worked on cellular images as part of the NeurIPS 2019 competition on disentangling biological signals from experimental noise for classifying gene mutations from cell images. Was in the top 5 percentile out of 1100 participating teams.
• 5. Also explored distributed training techniques to speed up training and improve model performance. The work was published in the European Symposium on Artificial Neural Networks, Computational Intelligence and Machine Learning (ESANN) 2021.

Aug 2018 – Dec 2018 · 4 mos

• I was the Teaching Assistant for the Machine Learning course offered in Fall 2018 semester at IIIT Delhi taken by 130 students. I prepared and graded assignments, quizzes, and mentored projects for final and pre-final year students at IIIT Delhi.

Aug 2016 – Dec 2018 · 2 yrs 4 mos

• I worked at the Computer Vision and Machine Learning (CVML) Lab, and the Speech Recognition Lab at IIIT Delhi. My work at CVML lab involved working on deep learning based fine grained classification and co-segmentation approaches for fashion parsing. It involved improving clothing segmentation by 2% for visual fashion image and attribute search systems. Also, exploited the pose structure to enhance SoA fine-grained classification performance by 2-3% across standard FGVC datasets. Published 2 papers at ICIP 2018. Work was in collaboration with Staqu Technologies. At the Speech Recognition lab, I worked on disentangling factors of variation in speech and emotion classification in collaboration with Verisk Analytics.

Aug 2017 – Dec 2017 · 4 mos · Bengaluru Area, India

• 1. Created a highly optimized sensor-based gesture detection and recognition model for on-device deployment surpassing benchmarks for memory constraints and output latency. Tested by deploying model on laptop hardware. We tested our system on Google Maps and music system applications by providing gesture controls.
• 2. The project was a 2M dollar IP project using ultrasonic sensors for VR device and laptop interactions.

Dec 2015 – Jul 2016 · 7 mos · New Delhi Area, India

• 1. EmuARM Emulator: Worked on building an emulator for compiling, running and debugging programs written in ARM and Thumb assembly languages.
• 2. DRAM Simulator: Built a cycle accurate DRAM memory controller model. It leverages a combination of event driven and cycle based simulations, and the performance is at par with the state-of-the-art simulator, DRAMSim2. It is part of the architectural simulator Tejas at IIT Delhi.
• Currently deployed in IIT Delhi, IIT Madras, IIT Patna, University of Adelaide, and other institutions.