Raaghav Kapoor — ML Engineer

About

Who I am.

I'm a B.Tech student at MAIT, New Delhi, with a deep focus on machine learning and deep learning research. I believe in building things that matter from model architectures to real pipelines.

My recent work spans sparse convolutional neural networks for LHC jet classification, achieving 82.25% accuracy with ~2× FLOPs reduction at 50% sparsity.

When not training models, I'm lifting, reading, or overthinking system architectures.

Work

Selected Projects.

01 Featured

End-to-End Event Classification with Sparse Neural Networks

Applying sparse convolutional neural networks to LHC jet classification. Pruning experiments show ~2× FLOPs reduction at 50% sparsity with minimal accuracy loss. Proposal incorporates ROC-AUC as primary evaluation metrics.

PyTorch Sparse NNs Python ML4Sci Physics ML

View on GitHub →

Sparse CNN for LHC Jet Classification

Sparse convolutional neural network applied to particle physics jet data from CERN's LHC. Implements structured pruning to reduce computational overhead while maintaining classification performance across quark/gluon jet types.

TensorFlow Sparse NNs Python ML4Sci Physics ML

View on GitHub →

Leaf Disease Classifier

End-to-end deep learning pipeline for plant leaf disease classification using a custom CNN built with TensorFlow/Keras. Complete training and inference pipeline.

TensorFlow CNN Keras

View →

Planet Classifier

Deep learning model for classifying planets based on dataset features. Built with TensorFlow to handle multi-class astronomical classification tasks.

TensorFlow Astronomy Jupyter

View →

Sentiment Review Classifier

Text classification model using deep learning to classify reviews as positive or negative. Handles sequence data with LSTM / dense architectures.

NLP TensorFlow Text ML

View →

TensorFlow Learning Journal

Comprehensive notes and code written while learning TensorFlow and neural networks. Covers fundamentals to advanced architectures — a living research notebook.

TensorFlow Neural Nets Notebooks

View →

Toolkit

Skills & Tools.

Core ML / DL

TensorFlow Keras PyTorch Sparse Neural Networks CNNs LSTMs Model Pruning ROC-AUC Analysis

Languages & Scientific

Python NumPy Pandas Matplotlib Scikit-learn Jupyter C++ JavaScript

Infrastructure & Tools

Git GitHub Linux VS Code Google Colab AWS (learning) HTML/CSS

Research Domains

Physics ML Computer Vision NLP Particle Physics (LHC) Open Source Research

Contact

Let's
talk.

Open to research collaborations, internship opportunities, and interesting conversations about ML, physics, and systems. Don't hesitate to reach out.

Send Email

✉ kapoorraaghav@gmail.com ⌥ github.com/kapoorraaghav ◈ linkedin.com/in/raaghav-kapoor-a0804b377/ ◉ MAIT, New Delhi — B.Tech IT, 2029

Raaghav Kapoor.

Who I am.

Selected Projects.

Skills & Tools.

Let'stalk.

Raaghav
Kapoor.

Let's
talk.