Vikhyat Chauhan
vikhyatchauhan@vt.eduContact Me
About
I'm a Master of Science in Computer Engineering student at Virginia Tech , where my research focuses on creating brain-inspired reinforcement learning frameworks for autonomous UAVs. My experience spans both industry and entrepreneurship; as a Software Engineer at GE Healthcare, I improved medical imaging diagnostics by integrating transformer-based vision models and redesigning MRI deployment with Kubernetes microservices. Previously, I co-founded TNM Electronics, where I served as CTO, architecting a patented, microservices-based IoT platform on AWS that scaled to over 100 devices. I am passionate about solving complex, end-to-end engineering problems, from developing ML inference services to building GPU-accelerated training pipelines.
Skills
Languages
Frameworks
Tools
Libraries
Experience
Graduate Research Assistant
Virginia Tech, VA
Brain-inspired Computer Architecture- National Science Foundation Grant
Nov 2024 - Present
- Designed HYDRA, a brain-inspired reinforcement learning framework for UAVs with constraint-driven arbitration (time/energy/space), integrating hybrid controllers. Built using PyTorch + Gym for training, Gazebo-ROS-PX4 for UAV simulation, and PID baselines for stability assessment.
- Implemented SAC with reward shaping and continuous control, leveraging GPU-accelerated training pipelines and simulation-to-real transfer workflows to enable robust, RL UAV decision-making under constrained environments.
GE Healthcare
Bangalore, IN
Software Engineer
Mar 2022 - Jun 2024
- Advanced ultrasound segmentation using SonoSAMTrack / SonoSAMLite, integrating transformer-based vision models (ViT, Swin), optimized with TensorRT for real-time clinical inference, improving throughput by 20%.
- Redesigned MRI deployment by migrating legacy C++ monolithic application to Kubernetes microservices, integrating MRPasting algorithm, and delivering optimized models on GE Edison with Docker and CI/CD (Jenkins, Git, SonarCloud) into PACS/RIS, reducing deployment time by 80% and enabling scalable rollouts.
TNM Electronics
Delhi NCR, IN
Co-Founder & CTO
Sep 2021 - Mar 2022
- Architected a microservices-based IoT platform on AWS EC2, scaling to 100+ devices & 1K+ daily events with RabbitMQ, MongoDB, and a custom MQTT broker on in-house patented hardware (achieved 99.9% uptime).
- Enabled predictive automation by embedding a Kafka pipeline in RabbitMQ to run federated learning models on PyTorch, delivering real-time user behavior analysis and privacy-preserving personalization.