Shravan S Rai

Explore my work in Robotics, AI and IoT for Telehealth

Education

Westcliff University, Irvine, California May 2024 - May 2026
Master of Science in Web Development & Health Informatics GPA: 4.00/4.00
Courses:

  • Big Data Analytics and Visualization
  • Managing Information Systems

Arizona State University, Tempe, Arizona Aug 2019 - May 2021
Master of Science in Robotics & Autonomous Systems GPA: 4.0/4.0
Courses:

  • Vehicle Dynamics & Control
  • Multivariable Controls
  • Modelling & Control of Robots
  • Computer Vision & Pattern Recognition
  • Robot Learning
  • Optimal Control & Reinforcement Learning
  • Linear Systems Theory

The National Institute of Engineering, Karnataka, India Aug 2013 - Aug 2017
Bachelor of Engineering Electronics & Communication Engineering GPA: 4.0/4.0
Courses:

  • Robotics
  • C++ and Data Structures
  • Control Systems
  • Operating Systems
  • Computer Vision & Image Processing

Industry Experience

Software Engineer III, Teladoc Health, Santa Barbara, California (June 2021 - Present)
- Benchmarking the latest SLMs with vision capabilities to deploy on the edge, leveraging AI+PC capabilities with NPU to ensure efficient operation of SLMs on edge devices.
- Developing and implementing a containerization strategy for models performing specific functions, creating a robust deployment system for edge devices using Azure IoT.
- Enhancing user engagement for a food logging application by integrating GPT vision capabilities and RAG to streamline and personalize user interactions.
- Spearheading the development & implementation of an AI-powered Virtual Sitter application that will be deployed to prevent patient falls in more than 150,000 hospital rooms.
- Devised an effective patient fall prevention solution by integrating multiple off-the-shelf Computer vision models, by rigorous benchmarking of multiple models to fit the requirements of fall prevention.

Robotics & Innovation Engineer II, Teladoc Health, Santa Barbara, California
- Leading the implementation efforts on the Autonomous Mobile Robot Navigation Software project.
- Enhanced the robot's mapping capability by transitioning from a Rao-Blackwellized Particle Filter-based SLAM algorithm to a factor-graph based SLAM algorithm.
- Implemented an A* based local planner for improved obstacle avoidance and navigation, optimizing the robot's pathfinding capabilities.
- Actively collaborated with team members to develop simulation software for testing and debugging the autonomous navigation of the robot, ensuring reliable performance.

Robotics Researcher Intern, Teladoc Health, Santa Barbara, California (June 2020 - April 2021)
- Crafted a state-of-the-art human tracking system utilizing deep learning models, capable of detecting a person within 20ms and maintaining efficient tracking within a 1.2-meter distance for person-following capabilities for a Telehealth Virtual Presence Robot.
- Devised a person reidentification algorithm by leveraging existing models to ensure continuous person tracking during the following process.
- Seamlessly integrated the person-following system into the existing autonomous navigation stack, resulting in a fully autonomous following robot.

Research Engineer, Motor Control, Robert Bosch, Bangalore, India (Sept 2017 - July 2019)
- Deployed Artificial Neural Network in the Control Algorithm for Switched Reluctance Motor Control on a Xilinx Zynq Ultra Scale + achieving computations in less than 5us.
- Developed Virtual System Model (Digital Twin) of Switched Reluctance Motor and its Control in MATLAB/Simulink, aiding the rapid prototype development as part of the R&D project.
- Developed Field Oriented Control algorithm for Permanent Magnet Synchronous Motor and BLDC motor for automotive applications.

Projects

Auto Rounding for Virtual Nurse (July '24)
Developed a Vision, SLM/LLM and Voice based application to perform patient rounding using telehealth devices installed in the room. The application surveys the patient room, extracts vital information, IV levels, Folley bag levels and other rounding information and updates the patient dashboard with the supervision of a virtual nurse. (patent pending)

Open-Source Developer: MRPT-ROS-PACKAGE (ROS1, ROS2, C++ , Python) (Apr '23 - present)
Developing advanced auto-navigation systems, focusing on efficient localization, dynamic path planning, and sophisticated behaviour planning. Proven track record in enhancing robotic autonomy and navigation accuracy in complex environments using ROS1 and ROS2. [Github]

BiteScan - Chat based AI Nutritionist (July '23)
Developed Contextual Nutritional Assistant using OCR and AI-based Chat Interface during a two-day hackathon. It comprises a mobile/web app providing personalized, real-time nutritional advice. Utilized OCR to interpret food labels from user-submitted images, coupled with user health data to offer custom advice. Delivered a unique solution addressing the gap in the market for instantaneous, personalized nutrition guidance.

Mobile Robot Autonomous Navigation (ROS, C++, Python) (Apr '23- Jun '23)
Achieved 3rd place in both simulation and real-world rounds at the ICRA- 23's BARN Challenge for Autonomous Mobile Robot Navigation, outperforming several global teams and demonstrating excellence in robot navigation.

Human Pose and Object detection ROS Node for Greenhouse Cobot (ROS2, C++ ) (Feb '23- Mar '23)
Collaborating with Dr. Jose Blanco from University of Almeria, worked on developing human pose detection ROS2 node and Fruit detection ROS2 node for a greenhouse robot using TensorRT framework on Nvidia Jetson TX2.

Technical Skills

Publications

Autonomous Ground Navigation in Highly Constrained Spaces: Lessons learned from the 2nd BARN Challenge at ICRA 2023

Autonomous Mobile Robot Obstacle Avoidance with Reinforcement Learning

Medium Articles:

Achievements & Awards