About

Sunil Patel is a highly experienced data scientist with over 8 years of relevant experience in the field. He is currently working as a Data Scientist - IV Deep Learning at Nvidia, where he is involved in large scale vision projects deployment with Nvidia TensorRT (C++ & Py), Deepstream (C++ & Py), and Triton. He has also worked as a Senior Solutions Architect - Deep Learning at Nvidia in the past. Sunil has 4 patents and 5 publications in the field of Deep Learning, which showcase his expertise in the area. He has also worked on 10+ projects and holds 10+ certifications in various technologies such as OpenCV C++, GStreamer, Boost Compute C++, and Thrust C++. He is proficient in Cuda Computing and has worked on improving efficiency by explicit memory assignment, partitioning schemes, compressing data, and utilizing Unified memory APIs for faster prototyping. Sunil is the creator of Aircontrol - an Open Source, Modular, Cross-Platform, and Extensible Flight Simulator For Deep Learning Research. He has deployed dockerized applications over Kubernetes and developed applications that take advantage of multiple GPU-multiple node environments. Sunil has a Master of Technology degree in Information Technology from iiit and is proficient in Deep Learning, Research Scientist, Data Scientist, OpenCV, Kubernetes, Pytorch, Computer Vision, Object detection, Object Detection, Tensorflow, and Senior Software Engineer. He has also given talks on various topics such as Building Indic ASR Using NVIDIA NeMo and Deploying Models Using Jarvis and Accelerated Database Query Using GPU. Sunil's past experiences include working on Unreal Engine, Autonomous Driving, Computational linguistics, and GPU Acceleration. He has worked on CUDA Computing, Nvidia DeepStream, and Nvidia Tensorrt. He has also developed models training NEMO API for Quartz-Net and Jasper with lightning-PyTorch and optimized them using Tensorrt and deploying them using the triton inference server. He has supported languages such as Hindi, Gujarati, and Punjabi.