I'm passionate about Machine Learning, Data Science, Large Language Models (LLMs) and Generative AI
I hold a Master’s degree from IIT Bombay and have over 2.3 years of experience in building and deploying machine learning and deep learning models.I also possess strong practical and theoretical expertise in the development of Large Language Models (LLMs) and Generative AI.
As an Analyst at D.E. Shaw India Private Limited, I have contributed significantly to leveraging machine learning and advanced analytics to drive impactful results. My role involved designing and validating predictive models using techniques such as Gradient Boosting, Random Forest, and Deep Neural Networks (DNNs), achieving impressive outcomes, including a 0.43 W/m² RMSE for irradiance prediction. I developed and implemented automated workflows, achieving a 95% reduction in data processing time, and applied advanced imputation techniques like Linear Regression, KNN, and MICE, improving data quality by 20%, which directly enhanced model performance.
I also spearheaded the extraction of actionable insights from over 500+ contracts through intelligent document analysis powered by Large Language Models (LLMs), significantly reducing manual review time by 80% and enhancing data accuracy by 45%. Additionally, I focused on building robust data validation frameworks, ensuring reliability and scalability across projects.
My work at D.E. Shaw reflects a strong commitment to integrating machine learning and automation into complex workflows, driving both efficiency and accuracy, while collaborating closely with cross-functional teams to achieve measurable business outcomes.
- Programming Languages: Python
- Machine Learning: Supervised, Unsupervised Learning, Time Series Forecasting,
- Deep Learning: CNN, LSTM, ResNet, Transformers,NLP
- Tools & Frameworks: PyTorch, TensorFlow, Scikit-learn, Pandas, NumPy, Matplotlib
- Database & Tools: SQL, Docker, Git,Flask
- Cloud: AWS,
I encourage you to explore my Machine Learning and Deep Learning projects. The links to the repositories are provided below, along with detailed descriptions at the bottom of this profile.
-
Customer Churn Prediction System
Developed an end-to-end ML pipeline for customer churn prediction, achieving an 83% ROC-AUC score, incorporating techniques like SMOTE for imbalanced data handling. -
Solar Power Forecasting Tool
Built and deployed a production-ready forecasting system using SVM, LSTM, and CNN-LSTM models, improving RMSE by 20% through rigorous validation protocols. -
MUSIFY - Music Composition using AI
Designed an AI-powered music composition system using piano notes to generate new musical pieces with minimal human intervention. -
Mango Leaf Disease Classification
This project utilizes deep learning techniques, particularly Convolutional Neural Networks (CNNs), to classify mango leaf diseases into 8 distinct classes. The goal is to improve disease identification and management in mango crops, ultimately boosting agricultural yield. -
Clothing Similarity Search
This project leverages SentenceTransformer('all-MiniLM-L6-v2'), a pre-trained model based on the MiniLM architecture, to compute dense vector representations of clothing item descriptions. The goal was to build a system that ranks similar clothing items based on text-based descriptions. -
Invoice Automatic Scanning and Text Predictions
This project automates the process of invoice reconciliation by detecting and extracting critical information such as Invoice Number, Billing Date, and Total Amount using YOLOv4 for object detection and Tesseract OCR for text recognition.Business Overview:
Traditionally, reconciling digital invoices is a manual, time-consuming task. This project aims to automate this process, saving significant time and effort across industries that rely on invoice processing.