Sushil Bharati Email & Phone Number

California, United States

Purchase, NY, US

• Scaled member-facing NLP-powered chatbot to support over 100 million members, enhancing user engagement and satisfaction.
• Collaborating with Hospitals and Health Systems (HHS) on innovative AI/ML projects in partnership with Microsoft and NVIDIA, driving cutting-edge healthcare solutions.
• Developing next-generation Generative AI Large Language Models for healthcare applications, pushing the boundaries of AI technology in the industry.

Purchase, NY, US

• Collaborated with Hospitals and Health Systems (HHS) to develop and implement Virtual Sitter with AI capabilities, from ideation to production-ready components.
• Awarded the Innovation Leadership Award 2023 (17.8% win rate) for pioneering AI/ML innovations in healthcare. Our team's Generative AI tech stack also won the People's Choice Award at the Summer Hackathon 2023.
• Advised on key AI/ML-based product innovations across Teladoc core teams and HHS teams, driving forward-thinking solutions.
• Led the migration of AI/ML tech stack to the Azure Cloud Platform, reducing costs by 300% through an efficient transition from a third-party ML tech vendor to Azure cloud.
• Trained, fine-tuned, and evaluated Language Models (LMs) using LoRA and other PEFT techniques on extensive text data (GBs), enhancing the scalability of member-facing chatbots within Microsoft Azure Infrastructure.
• Designed and deployed an end-to-end MLOps pipeline to support label annotation, training language models, real-time production data analytics, train/test metrics evaluation, automated deployment (CI/CD), and production.

Purchase, NY, US

• Led team to increase the virtual Health Assistant interaction satisfaction by 25% via conceptualizing and designing/maintaining internal toolboxes for monitoring, assessing, and improving the deployed NLP model.
• Led capstone team project to a second-place finish at the University of California, SantaBarbara in collaboration with Teladoc to train, test, and deploy Machine/Deep Learning models for assessing and visualizing.

Purchase, NY, US

• Improved and scaled existing Health Assistant (Medical Dialogue System) using Natural Language Processing models/algorithms and data-driven logistics by working closely with the diverse teams across the organization to.
• Advised Robotics intern for the "Follow Me" project which spans internal framework design decisions, extension, and integration with existing RP VITA™ robot app and InTouch Provider™.

• Designed real-time InTouch Nametag™ pipeline from conception to rollout among Lite4™ and Vici-Next™ InTouch devices by overcoming challenges such as AI biases, robustness together with optimizations to run real-time on.
• Designed deep learning based model for 'Human Sentiment Analysis' to support behavioral health analysis and improved the accuracy by 8% over "Training Deep Networks for Facial Expression Recognition with Crowd-Sourced.
• Collaborated with the students in their senior year at the University of California SantaBarbara for a Capstone Project that span Human-Computer Interaction, UI/UX Design and Artificial Intelligence for improved.

• Designed Web-VR pipeline to convert 360 video to virtual tour for medical research/educational purposes.
• Innovated novel pipeline for robust key object segmentation in video-sequence (skipped frames).
• Successfully tested 3D reconstruction techniques for featureless objects (small/large).
• Supervised intern for non-AI/ AI-based human skin segmentation project plus testing/debugging.
• Assisted CEO for proposal/grant writing.
• Worked closely with CTO to revamp the existing code-structure to migrate to the production phase.

Lawrence, KS, US

• Usually, salient object detection and tracking is vital to unsupervised detection of a moving object via drones or flying UAVs but consumes heavy computing time. Thus, I was involved in research to build efficient.
• Conducted NASA-funded research on Sense-and-Avoid Collision for Unmanned Aerial Vehicle (UAVs).
• Worked on Computer Vision Application - Salient Object Detection and Object Tracking.
• Published "Fast and Robust Object Tracking with Adaptive Detection" in 28th IEEE International Conference on Tools with Artificial Intelligence (ICTAI). (acceptance rate: 31%)[Paper.
• Published "Real-time Obstacle Detection and Tracking for Sense-and-Avoid Mechanism in UAVs" in IEEE Transactions on Intelligent Vehicles. [Project page: https://www.ittc.ku.edu/cviu/tracking.html] [Paper.
• Published "RES-Q: A Robust Outlier Detection Method in Stereo Images for Fundamental Matrix Estimation" in IEEE Access. (impact factor: 3.557)[Paper: https://ieeexplore.ieee.org/document/8451885]

Lawrence, KS, US

• Courses taught:
• EECS 168 Software Design, Algorithm and C++ (Fall 2017)
• EECS 318 Circuit and Electronics Lab (Spring 2016, Spring 2017)
• EECS 138 Introduction to Computing - Matlab (Fall 2016)

Lawrence, KS, US

• Conducted Research on Object Detection and Tracking for autonomous Unmanned Aerial Vehicles (UAVs).
• Performed pre-deployment tracking algorithmic tests on NVIDIA Jetson TK1 (light-weight embedded system).
• Implemented a novel approach to fast detect and track aerial objects for long-term tracking than current state-of-the-art methods.

• Worked with Rapid Innovation and Development (RIDL) department on 'Computer Vision', 'Deep Learning', and 'Thermal' applications in the company’s RP- VITA™ robots.
• Implemented the company’s VITA™ robot usability for the person following in a clinical setup by application of deep learning and computer vision algorithms.
• Used PrimeSense RGB-D camera along with OpenNI libraries for depth estimation and processing in VITA™ robots.
• Advanced thermal BOSON™ camera applications in VITA™ robots using methodical research techniques [side project].
• Conducted experiments to test the reliability of Philips licensed non-contact heart rate monitoring algorithm in an assigned rPPG (remote PhotoPlethysmoGraphy) project [side project].My roles (main project):+.

Kathmandu, NP

• Trained wide range of programming languages – C/C++, MATLAB (Image Processing), Python.
• Trained and developed live projects and commercial websites using PHP, SQL, HTML (5.0), CSS (3.0) and JS.
• Drove live project management through a strong focus on customer service, merchandising, and teamwork

• Trained wide range of programming languages – C/C++, MATLAB (Image Processing), Python.
• Trained and developed live projects and commercial websites using PHP, SQL, HTML (5.0), CSS (3.0) and JS.
• Drove live project management through a strong focus on customer service, merchandising, and teamwork

Ms In Ee (With Honors), Focus Area: Computer Vision

Nanodegree, Machine Learning

Nanodegree, Deep Learning

Nanodegree, Deep Reinforcement Learning For Enterprise

Bachelor'S Degree, Electronics And Communications Engineering

Slc