Research Experiences
Research Scientist | Rockwell Automation
July 2023 - Present | Austin, TX, USA
- Designed LenzAI, a Closed-Loop Machine Vision AI solution for seamless integration by industrial operators.
- Advanced AI-driven data labeling platform utilizing Instance Segmentation (SAM) and Monocular Depth Estimation (DINOv2), reducing per-image annotation time by 10X
- Integrated Active Learning with Model-Health Monitoring to minimize redundant labeling and alert for False Positives in industry settings
- Developed a novel zone transformation and unsupervised depth-based wrinkle detection system for automotive seat inspection.
- Proposed a novel video-processing AI for CPU/GPU to model periodic patterns, enabling real-time closed-loop control systems for dynamic state classification and actuation
Graduate Research Assistant | University of Michigan
Jan 2022 - May 2023 | Ann Arbor, MI, USA
- Proposed Multi-Mode Video Compression (MMVC) framework with block-based prediction mode selection and adaptive entropy coding, CVPR 2023
- Outperformed SOTA learning-based and conventional codecs on popular benchmark datasets in PSNR & MS-SSIM metrics, specifically a 1dB PSNR and 0.02 MS-SSIM improvement at a very low, 0.1 bit-rate.
Computer Vision Engineer | Toshiba Software Pvt
July 2019 - Aug 2021 | Bangalore, KA, India
- Designed an Item Recognition software for No Touch Checkout, enabling automated billing in grocery stores
- Deployed solution achieved a 98% classification accuracy in real-time speed.
Undergraduate Research Assistant | VISA Lab
July 2018 - June 2019 | Tirupati, AP, India
- Developed a novel end-to-end deep learning model to profile the 3D shape of objects from deformed fringe patterns, CVIU 2020
- Reconstructed shapes at -5dB SNR with a 0.005 RMSE surpassing traditional approaches by 10 folds.