Work Experience
Aug '21 - Present
Sr. Applied Scientist
@ Amazon Web Services (Sunnyvale, CA)
• Led development of AI training & inference pipelines leveraging LLMs and VLMs to enable scalable AI solutions for AWS customers; Optimized multi-modal training pipelines using FSDP & FP8, and deployed models on humanoid robots, enabling more efficient edge inference.
• Developed and owned GenAI@Edge product offering that enabled over 10 customers to deploy CVML solutions at the edge; Successful deployments saved customers an average of $20M in automation costs and drove over $5M in average annual recurring revenue through adoption of AWS services.
Nov '19 - Aug '21
Sr. Computer Vision Engineer
@ Dolby Laboratories (Sunnyvale, CA)
• Led design, optimization, and deployment of computer vision and deep learning pipelines for Dolby ATG and Dolby iAPI products, enabling real-time video and image processing applications.
• Developed multimodal pipelines supporting segmentation, classification, object detection, GANs, and pose estimation to enhance
Dolby's computer vision capabilities.
• Optimized model deployment on embedded devices, improving inference latency by nearly 400% to 25ms to support low-latency
applications.
Dec '18 - Dec '19
CoFounder and CTO
@ Strada Laboratories (San Francisco, CA)
• Co-founded an urban analytics startup applying computer vision to analyze city movements and assist urban planning and development, hereby automating the process & saving costs for city planning.
• Raised seed funding through Y-Combinator; developed optimized cycle lane solution integrated in Chinatown, San Francisco.
Jan '18 - Dec '19
AI Research Engineer
@ Ford Motor Company (Palo Alto, CA)
• Drove AI innovation for autonomous and connected vehicles.
• Enhanced CV solutions for Ford Performance Racing (NASCAR) using deep learning for pre-race, in-race, and post-race analysis.
• Published research applying CV, ML, DL and RL to mobility domains.
Jan '17 - Aug '17
Computer Vision Research Engineer (Intern)
@ Volvo (Shippensburg, PA)
•
Developed and optimized object detection and tracking solutions on embedded systems for semi-autonomous construction
vehicles, improving latency, detection accuracy.
•
Integrated and fused camera, RADAR, and stereo vision data to enhance perception capabilities.
May '14 - Aug '15
CoFounder and CTO
@ Tellmate Helper (Ahmedabad, India)
• Co-founded and led the making of ‘Tellmate’, a device made using Kinect360 and Intel RealSense camera integrated with PandaBoard ES for assisting visually impaired people; Selected for Top 20 startups in India from 1.9k participants; 200k INR seed funding by Intel Digital India Challenge 2015.
Education
Sep '15 - Dec '17
MS in Computer Engineering
@ Northeastern University (Boston, MA)
Jul '10 - May '14
BTech in Electrical Engineering
@ Nirma University (Ahmedabad, India)
