Skip to content
View Mohithavelagapudi's full-sized avatar
👋
👋
5 followers · 0 following

Block or report Mohithavelagapudi

Block user

Prevent this user from interacting with your repositories and sending you notifications. Learn more about blocking users.

You must be logged in to block users.

Maximum 250 characters. Please don't include any personal information such as legal names or email addresses. Markdown supported. This note will be visible to only you.
Report abuse

Contact GitHub support about this user’s behavior. Learn more about reporting abuse.

Report abuse
Mohithavelagapudi/README.md

🚀 About Me

package main

type Developer struct {
	Name      string
	Title     string
	Company   string
	AlmaMater string
	City      string
	Passions  []string
	Motto     string
}

func main() {
	me := Developer{
		Name:      "Mohitha Velagapudi",
		Title:     "AI & Software Engineer",
		Company:   "Guidewire Software",
		AlmaMater: "Amrita University",
		City:      "Bengaluru, India",
		Passions:  []string{"Machine Learning", "Data Science", "AI Research", "Scalable Systems"},
		Motto:     "Turning data into decisions, and ideas into impact.",
	}
}

💡 "Data is the new oil, but it's AI that's the combustion engine." - My philosophy on modern tech

📚 Coursework

📘 This represents the in-depth coursework syllabus covered throughout my undergraduate studies

🔬 Featured Projects

Custom GPU Operators for Deep Learning

High-performance Triton-based GPU kernels accelerating core deep learning ops — from matrix multiplication to convolutions and activations.

Impact: 3× faster training on custom kernels

🧠 Built from First Principles

Bite2Burn: AI-Powered Calorie Estimation & Nutrition Assistant

RAG-based nutrition assistant built over Indian Food Composition Tables (IFCT 2017), combining NLP, retrieval, and nutrition science for culturally-aware calorie estimation.

Impact: AI-driven dietary guidance for 100+ regional cuisines
🧠 Integrates LLMs, RAG, ChromaDB, OCR & structured nutrition data

Vision-Based Gesture-Controlled Robot

End-to-end robotic control system integrating computer vision, deep learning, and embedded motor control for real-time gesture-driven navigation.

Impact: Achieved 95% gesture recognition accuracy on live video

💡 Human-robot interaction via vision-based control

Unsupervised Occlusion-Aware Image Animation

Unsupervised framework for realistic image animation using learned keypoints, local affine motion, and occlusion-aware generation.

Impact: Achieved temporally coherent animations without pose supervision
🧩 Combines keypoint learning, motion modeling & occlusion-aware synthesis

📝 Research & Publications

📄 Audio Driven Detection of Hate Speech in Telugu: Toward Ethical and Secure CPS

Authors: Ajay Surya Jampana, Mohitha Velagapudi, Santhosh Kumar M*, Sai Ravula P*, Prasanna Teja M, Jyothish Lal G

Affiliation:
School of Artificial Intelligence, Amrita Vishwa Vidyapeetham,Coimbatore, Tamil Nadu, India

Conference: International Conference on Cyber-Physical Systems (CPS) for Sustainable Development, 2024 Published by: Springer Nature Year: 2024

🔗 Access Paper

ACL Anthology PDF

📈 Impact Visualization


Model Performance
Accuracy
████████████░░ 79%  Multi class
███████████████ 89%  Binary class

📄 Exploring Kolmogorov Arnold Networks for Interpretable Mental Health Detection and Classification from Social Media Text

Authors: Ajay Surya Jampana, Mohitha Velagapudi, Neethu Mohan, Sachin Kumar S

Affiliation:

Amrita School of Artificial Intelligence, Coimbatore, Amrita Vishwa Vidyapeetham, India

Conference: 21st International Conference on Natural Language Processing (ICON) 2024
Published by: NLP Association of India (NLPAI)
Location: AU-KBC Research Centre, Chennai, India
Pages: 206–214 Year: 2024

🔗 Access Paper

ACL Anthology PDF

📈 Impact Visualization

Model Performance
█████████████░ 93% Accuracy
█████████████░ 91% F1-Score
████████████░░ 85%
Interpretability

📄 Enhancing Power Quality Disturbance Classification through Ensemble Learning and Statistical Techniques

Authors: Mohitha Velagapudi, Ajay Surya Jampana, Neethu Mohan, Sachin Kumar S, Soman KP

Affiliation:
Amrita School of Artificial Intelligence, Coimbatore, Amrita Vishwa Vidyapeetham, India

Conference: 2024 10th International Conference on Advanced Computing and Communication Systems (ICACCS)
Published by: IEEE
Location: Coimbatore, India

🔗 Access Paper

IEEE Xplore PDF

📈 Impact Visualization

Model Performance
█████████████░ 98.4% Accuracy
█████████████░ 98.4% F1-Score
█████████████░ 98.72% 10-CV

📄 YOLOv10 for Enhanced Trypanosome Detection

Authors: Mohitha Velagapudi, Ajay Surya Jampana, Sowmya V, Vinayakumar Ravi

Affiliations:
1️⃣ Amrita School of Artificial Intelligence, Coimbatore, Amrita Vishwa Vidyapeetham, India
2️⃣ Center for Artificial Intelligence, Prince Mohammad Bin Fahd University, Khobar, Saudi Arabia

Conference/Book Chapter: Decoding Medical Images: Enhancement, Restoration, Reconstruction
Published by: River Publishers
Year: 2024

🔗 Access Chapter

River Publishers
PDF

📈 Impact Visualization

Model Performance
███████████░░░ 79% Precision
█████████░░░░░ 72% Recall
██████████░░░░ 76% F1-Score

📈 Coding Stats

Python       12 hrs 45 mins  ████████████░░░░░░░░░   48.2%
Go      	 8 hrs 30 mins   ████████░░░░░░░░░░░░░   32.1%
YAML         2 hrs 15 mins   ██░░░░░░░░░░░░░░░░░░░    8.5%
SQL          1 hr 45 mins    █░░░░░░░░░░░░░░░░░░░░    6.6%

Last Updated: October 2025 | Built with 💙 and lots of ☕

Pinned Loading

  1. Audio-Driven-Hate-Speech-Detection-in-Telugu Audio-Driven-Hate-Speech-Detection-in-Telugu Public

    Low-resource multimodal hate speech detection leveraging acoustic and textual representations for robust moderation in Telugu.

    Jupyter Notebook

  2. Bite2Burn-AI-Powered-Calorie-Estimation-Personalized-Nutrition-Assistant Bite2Burn-AI-Powered-Calorie-Estimation-Personalized-Nutrition-Assistant Public

    An AI + RAG system delivering culturally contextualized nutritional intelligence, personalized caloric guidance, and queryable knowledge over Indian Food Composition Tables (IFCT).

    Jupyter Notebook 1

  3. Custom-GPU-Operators-for-Deep-Learning Custom-GPU-Operators-for-Deep-Learning Public

    High-performance Triton-based GPU kernels for accelerating core deep learning operations, from matrix multiplication to convolutions and activation functions.

    Python

  4. Real-Time-Image-Animation-via-Learned-Keypoints-and-Local-Affine-Motion Real-Time-Image-Animation-via-Learned-Keypoints-and-Local-Affine-Motion Public

    Unsupervised deep framework that animates a source image using motion extracted from a driving video through learned keypoints, local affine transforms, dense motion fields, and occlusion-aware gen…

    Python

  5. RL-Algorithms-From-Scratch-With-GYM-Environments RL-Algorithms-From-Scratch-With-GYM-Environments Public

    Derived and Implemented 9 RL algorithms from scratch in Python and simulated with GYM environments

    Python 1

  6. Vision-Based-Gesture-Controlled-Robot-using-Deep-Learning Vision-Based-Gesture-Controlled-Robot-using-Deep-Learning Public

    An end-to-end robotic control system built from scratch — integrating computer vision, deep learning, and motor control for real-time gesture-driven navigation.

    Python