ML/AI · Computer Vision · AI Agents

Building things, breaking things, and learning why they broke.

I'm a B.Tech Mechanical Engineering student at MNIT Jaipur, exploring Machine Learning and Computer Science alongside my core engineering background.

My current interests include:

• 🤖 Machine Learning & Artificial Intelligence
• 👁️ Computer Vision
• 🧠 Vision-Language Models
• 🔬 ML Research
• 🌱 Open Source
• 🛠️ Building practical ML applications

Currently working on Vision-Language Models for automated chest X-ray report generation and exploring how vision and language models can work together for real-world problems.

VLM · PyTorch · Hugging Face Transformers

Research project at MNIT Jaipur.

Currently reproducing R2GenGPT on the IU X-Ray dataset to establish a Vision-Language baseline for automated radiology report generation.

Working on:

• 📷 7,470 X-ray images and 3,955 reports
• 🧠 ViT vision encoder + Llama 2 7B decoder
• 🔗 Projection layer connecting visual and language representations
• 🏥 Image-report pair training
• 📊 Clinical evaluation using GREEN and RadGraph F1
• 📈 Comparison with BLEU, ROUGE-L, METEOR and CIDEr

Python · Flask · Scikit-learn · Pandas · NumPy

A machine-learning powered CGPA prediction system built from student performance data.

• Engineered 15 features from 370+ survey responses
• Combined Random Forest, Gradient Boosting, Extra Trees and Ridge into a weighted ensemble
• Achieved 60%+ within-one-grade accuracy on unseen data
• Implemented k-NN SMOTE from scratch
• Built and deployed a Flask REST API
• Provides subject-level predictions and semester forecasting

Python · TensorFlow · Keras · OpenCV

A real-time computer vision system for detecting face masks.

• Trained a CNN on 3,800+ images
• Applied data augmentation during training
• Achieved 95%+ detection accuracy
• Implemented real-time inference using an OpenCV webcam feed

Contributed to open-source projects across Agentic AI and accessibility engineering.

• 🔀 Merged 10+ PRs across 4 repositories
• 🤖 Built a news-summarizer-agent for fetchai/innovation-lab-examples
• 📰 Implemented an agentic RAG system using NewsAPI and ASI
• 🔐 Fixed an OTP spam-prevention issue
• ⚡ Resolved a real-time Socket synchronization bug

• 🏅 Rank 3672 in ICPC Regional among 13,000+ participants nationwide
• 📈 98.06 percentile in JEE Main among 14.2 lakh candidates
• 🏆 20th in Uttar Pradesh in NTSE Stage 1
• 🥉 3rd zonally in the International Mathematics Olympiad

Malaviya National Institute of Technology, Jaipur

B.Tech in Mechanical Engineering
2024 – 2028

• 🧠 Multimodal AI — connecting vision and language
• 👁️ Computer Vision — understanding images beyond classification
• 🔬 ML Research — reproducing papers and experimenting with architectures
• 🤖 Agentic AI — building systems that can reason and act
• 🌱 Open Source — contributing, debugging, and building with others