Romil Shah Email & Phone Number

San Jose, Ca, Us

San Jose, California, United States

• Lead and manage development of GenAI training and inference pipelines for LLMs, VLMs, and LBMs• EdgeFM Accelerator to scale FM inference for cheaper, faster and more efficient FM deployment

San Francisco, California, United States

• Create/Deploy/Manage Deep Learning pipelines for Cloud and IoT Edge; Design AWS architecture for Edge IoT based ML solutions• Lead multiple engagements to solve CVML related customer requirements; Dive deep to improve AI/ML AWS Services• Own and Manage customized Edge ML pipelines; Develop MLOps for customersKey Technologies: Amazon SageMaker, AWS IoT Core, AWS Greengrass, AWS Panorama, NVIDIA Jetsons, TensorFlow, PyTorch, OpenCV

Sunnyvale, California, United States

• Design/Optimize/Implement/Deploy CV + DL API pipelines for Dolby.IO and Dolby iAPI• Train/Deploy/Optimize DL models for various Dolby vision applications• Working on real time application for segmentation, object detection, GANsKey Technologies: OpenVino DLDT, Tensorflow, TF JS, ONNX Runtime, Chromium

San Francisco Bay Area

• Co-Founded the startup which focused on using Computer Vision for analyzing urban movements• Analyze patterns to aid urban planning for creating smarter and safer cities

Palo Alto, California, United States

• Worked on projects for Ford Performance NASCAR Racing teams to analyze images/videos/telemetry using CV/ML/DL• Projects included object detection, pose estimation, segmentation, SfM (Structure-from-Motion), DNN Frame interpolation, DNN motion estimation• 7 Patents Pending in Mobility/Automotive using AIKey Technologies: TensorFlow, PyTorch, OpenCV, dlib

Shippensburg, Pennsylvania

• Deep Learning for Autonomous Construction Equipment vehicles with L2/L3 capabilities• Improving state-of-the-art DNN architecture and optimize for sensor fusion, embedded systems with multi-communication nodes• DNN based Object detection, DNN based object tracking Key Technologies: OpenCV, DLib, ZMQ, CNN, Caffe/Caffe2, CUDA

Ahmedabad

• Contributed in making of ‘Tellmate’, a device for visually impaired people using Computer Vision algorithms• Selected for Top 20 startups in India by Intel Digital India challenge 2015Key Technologies: OpenCV, C++, Python, Kinectwww.tellmate.in

Boca Raton, Florida

Worked on Video processing, scene analysis under Dr H Kalva.Project was based on video processing and communication. Image Processing using the MPEG-2 codec as well as H264 standard was included.Compression of a video was done using X264 and FFMPEG libraries where the redundant pixels were analyzed, removed and the respective motion vectors were retained.

Master’S Degree, Computer Vision, Machine Learning And Algorithms, Ece Dept

Activities and Societies: Graduate Student CouncilTaking up courses to focus my masters more towards Computer Vision and Machine Learning..

Bachelor Of Technology (B.Tech.), Electrical, Electronics And Communications Engineering

Bachelors of Technology in Electronics and Communication Engineering Projects include: Worked on PIC, AVR and 8051 Microcontrollers, Intel.

12Th, Science, 94.3