I'm a Predoc Associate at the Indian Institute of Science (IISc), Bangalore, working at the AiREX Lab under Professor Sashikumaar Ganesan. My research focuses on developing cutting-edge Machine Learning systems for applications in combat aircraft.
I hold a B.Tech in Electrical Engineering from the National Institute of Technology, Hamirpur, where I cultivated a deep passion for Machine Learning, Optimization, and Natural Language Processing.
In the past, I’ve worked as:
- Data Scientist at Bajaj Finserv Health, where I built a high-performance statistical model for fraud detection.
- ML Intern at IIT Indore, where I collaborated with Professor Vimal Bhatia on stock price prediction models.
- Open Source Technical Writer at OpenMined, contributing to documentation and tutorials for differential privacy and federated learning.
A forgery detection system to classify alterations (e.g., overwriting, scribbling) using YOLOv8. Built an extensive dataset of 600+ document images and designed a Streamlit-based web interface for user-friendly analysis.
Developed an end-to-end system to detect ophthalmic and systemic diseases (e.g., Diabetic Retinopathy, Cataract) via smartphone cameras. Integrated deep learning with 3D-printed smartphone accessories for affordable healthcare.
Implemented ML models predicting stock prices with 85% accuracy, leveraging time series analysis and news data. Boosted model performance by 25% through advanced feature engineering.
- Languages: Python, C++, SQL, Solidity
- Libraries: TensorFlow, PyTorch, OpenCV, Scikit-learn, NumPy, pandas, YOLOv8
- Other Tools: Google Cloud Platform, Firebase, Neo4j, MATLAB, LaTeX
- Technical Writing: Published articles on federated learning, privacy, and ML at Medium and Hashnode.
- Contributing to PULSE (Python Unified Library for Sensor Emulation), an open-source library for radar signal processing and ML testing.
- Exploring Neural Radiance Fields (NeRFs) and their applications in 3D rendering.
- Preparing to join NLP Lab at IISc Bangalore as a predoctoral researcher.
Feel free to explore my work and open a discussion or collaboration! 🚀