Varun Ganjigunte Prakash

Senior Research Scientist — Computer Vision & Applied AI

Frontera · Bengaluru, Karnataka, India

CV

About

I am a computer vision engineer and researcher interested in building intelligent systems that have direct real-world impact. My work focuses on the intersection of computer vision, multimodal machine learning, and applied AI, with applications in healthcare technology and autonomous systems.

Currently, I work as a Senior Research Scientist at Frontera, where I build production computer vision and multimodal AI systems — object detection fine-tuning, video understanding for behavior recognition, document intelligence for clinical assessments, and speaker verification. Previously, I developed AI-powered assessment tools for children with autism spectrum disorder at CogniAble, and worked on computer vision applications in industrial automation at L&T Technology Services.

My research interests include computer vision, multimodal learning, action recognition, and machine learning for healthcare, alongside a continuing interest in robotics and deep reinforcement learning. I am particularly motivated by applications that can help people in meaningful ways — from assisting children with developmental challenges to making autonomous systems safer and more useful.

Research Interests

  • Computer Vision & Object Detection
  • Multimodal & Video Understanding
  • Action & Behavior Recognition
  • Document Intelligence & Vision-Language Models
  • Speech and Speaker Modeling
  • Machine Learning for Healthcare

Publications

Computer Vision-Based Assessment of Autistic Children: Analyzing Interactions, Emotions, Human Pose, and Life Skills

Varun Ganjigunte Prakash, Manu Kohli, Swati Kohli, Prathosh A. Prasad, Tanu Wadhera, Diptanshu Das, Debasis Panigrahi, John Vijay Sagar Kommu

IEEE Access, 2023

Paper PDF

Video-based real-time assessment and diagnosis of autism spectrum disorder using deep neural networks

Varun Ganjigunte Prakash, Manu Kohli, Prathosh A. Prasad, Monica Juneja, Manushree Gupta, Smitha Sairam, Sadasivan Sitaraman, Anjali Sanjeev Bangalore, John Vijay Sagar Kommu, Lokesh Saini, Prashant Ramesh Utage, Nishant Goyal

Expert Systems, Wiley Online Library, 2023

Paper

Deep Learning-Based Human Action Recognition Framework to Assess Children on the Risk of Autism or Developmental Delays

Manu Kohli, Arpan Kumar Kar, Varun Ganjigunte Prakash, Prathosh A. Prasad

International Conference on Neural Information Processing (ICONIP), CCIS, Springer, 2023

Paper

Behavior-Aware Robot Navigation with Deep Reinforcement Learning

Varun Ganjigunte Prakash

IEEE 6th International Conference on CSITSS, 2022

Paper

Autonomous Service Robot

Arshad Javeed, Varun Ganjigunte Prakash, Sudarshan Patilkulkarni

Advances in Robotics (AIR 2019), ACM ICPS, 2019

Paper

Selected Research & Projects

Behavior-Aware Robot Navigation with Deep Reinforcement Learning

Oct 2022 - Dec 2022

Implemented a human behavior-aware crowd navigation model utilizing deep reinforcement learning algorithms with enhanced risk awareness, safety, and understanding of the environment for robot navigational decisions.

Deep RL Robot Navigation Human-Robot Interaction

Computer Vision-Based Assessment of Autistic Children

IEEE Access, 2023

Developed deep learning models to analyze interactions, emotions, human pose, and life skills for assessment of children with autism spectrum disorder, enabling real-time behavioral analysis.

Computer Vision Healthcare AI Action Recognition

6-DOF Robotic Arm for Object Manipulation

Jun 2019 - Mar 2020

Designed a custom 4-axis robotic arm and implemented pose estimation, planning, and control in ROS. Tested the manipulator for 3D object grasping with deep learning-based perception.

ROS Manipulation Inverse Kinematics

Autonomous Service Robot (DEXTER)

AIR 2019 (ACM ICPS)

Designed a mobile manipulator robot to assist elderly and differently-abled persons with daily tasks. The system used voice commands and computer vision for object search and delivery in home environments.

Mobile Manipulation Voice Control Assistive Robotics

NLP Models for Contract Data Mining

Colt Technology Services, 2023

Implemented and integrated natural language processing models for contract data extraction and sentiment analysis, supporting business intelligence applications.

NLP Information Extraction Sentiment Analysis

Indian Currency Recognition System

2017

Developed a deep learning model for Indian currency recognition and deployed it as an Android mobile application to assist visually impaired users.

Mobile App Image Classification Accessibility

Experience

Senior Research Scientist

Frontera | Full-time, Remote

Jan 2025 - Present

  • Behavior recognition research across VideoMamba, HumanOmni, and pose pipelines
  • Speaker signature registration and verification (92.2% person-level accuracy)
  • Facilitated India-team stand-ups, code reviews, and ML design discussions; resolved blockers across time zones
  • RF-DETR fine-tuning for form-structure detection (0.936 AP50)
  • Multimodal document AI for clinical assessments (94-98% field accuracy)

Senior Computer Vision Engineer

Frontera | Full-time, Remote

Dec 2023 - Dec 2024

  • Production social-engagement vision: person tracking, child/adult detection, gaze models
  • Gaze and head-pose fusion to reduce false look-at predictions
  • Clinical-video ASR quality: ghost-transcript resolution and language detection

Senior Engineer, Data Science

Colt Technology Services | Full-time, Hybrid

Mar 2023 - Dec 2023 · 10 months

Bengaluru, Karnataka, India

  • Implemented and integrated NLP models for Contract Data Mining
  • Implemented NLP models for Sentiment Analysis
  • Supported ML models for Opportunity Prioritization

Head Machine Learning and Technology Architect

CogniAble | Full-time, Remote

Jul 2022 - Feb 2023 · 8 months

Remote, India

  • Researched and implemented computer vision approaches for clinical screening
  • Implemented neural network-based recommender system for treatment personalization

Machine Learning Engineer - Computer Vision

CogniAble

Mar 2021 - Jul 2022 · 1 year 5 months

  • Trained and deployed custom action recognition deep networks
  • Implemented ML classifiers for neurodevelopmental disorder treatment personalization
  • Led development of 9 computer vision models in production

Engineer

L&T Technology Services Limited

Jun 2018 - Mar 2021 · 2 year 11 months

Bengaluru

  • Developed computer vision applications for edge inference products
  • Software development in C++ and Lua for medical equipment
  • Implemented OCR pipeline for HMI screen automation (2x time savings)

Computer Vision Intern

Skylark Drones

Feb 2018 - May 2018 · 4 months

Greater Bengaluru Area

Technical Skills

Computer Vision

RF-DETR, YOLO-World, YOLOv8, SSD, action recognition, pose and head-pose estimation, person tracking, OpenCV, COCO evaluation

Multimodal & Speech

Gemini Vision, HumanOmni, InternVideo2, VideoMamba, V-JEPA, YAMNet, WeSpeaker, Wav2Vec2, pyannote, Silero VAD

ML Frameworks

PyTorch, TensorFlow, Keras, scikit-learn, Hugging Face Transformers, ROS

MLOps & Data

Temporal, AWS (S3, Batch), Docker, CUDA, Git, Weights & Biases, Label Studio, gold-set regression testing

Education & Certifications

Sri Jayachamarajendra College Of Engineering

Bachelor of Engineering - Electronics and Communication

2014 - 2018

Grade: GPA: 8.47/10.00

Activities: e-Yantra Robotics Lab

Certifications

  • Robotics Software Engineer - Udacity (Apr 2020)
  • AI Challenge Impact - Omdena (Mar 2019)
  • DELF B1 - Diplôme d'études en langue française (Mar 2017)

Contact

I'm always interested in discussing new research opportunities, collaborations, or interesting problems in computer vision and applied AI.

Email: varun.ganjigunteprakash@gmail.com

Location: Bengaluru, Karnataka, India

CV: CV

Links: LinkedIn · GitHub · Google Scholar