I'm a dedicated engineer passionate about harnessing Machine Learning, Data Science, Data Engineering, and Software Development to create impactful solutions. With a strong foundation in building scalable systems and extracting insights from data, I thrive on solving complex challenges.
Outside of coding, I enjoy exploring new tech trends and collaborating on innovative projects. Let’s build something extraordinary together!
Saayam For All | March 2025 - Present
Accenture | February 2022 - August 2023
Developed a robust and scalable social media REST API using FastAPI with comprehensive user authentication, content management, and social interaction features. The application implements a complete CRUD system for posts with JWT-based authentication, bcrypt password hashing, and role-based access control ensuring secure user operations. Built with PostgreSQL database integration using SQLAlchemy ORM and Alembic for database migrations, the API supports user registration, login, post creation/editing, and a voting system for social engagement. The entire application is containerized with Docker for seamless deployment across different environments, while comprehensive test coverage with pytest ensures reliability and maintainability. Interactive Swagger UI documentation provides effortless API exploration and testing, making it ideal for frontend integration and developer collaboration.
View Project →Developed a character-level bigram language model from scratch using Python and PyTorch to learn probabilistic relationships between characters in a text corpus. The model predicts the next character based on the current one and generates synthetic text sequences that mimic the statistical patterns of the training data. The training loop, loss computation, and text sampling were implemented manually to gain a deep understanding of language modeling fundamentals. This project showcases foundational NLP techniques and neural network training without relying on high-level frameworks or pretrained models.
View Project →Built a neural network to classify movie reviews as positive or negative using the IMDB dataset and Keras. The model processes pre-tokenized integer sequences representing reviews and learns sentiment patterns using an embedding layer followed by fully connected dense layers. Implemented custom decoding logic to map token indices back to readable text for interpretability. Trained the model on a vocabulary of the top 10,000 most frequent words, achieving high accuracy on unseen test data. This project demonstrates a practical application of deep learning in natural language processing using real-world sentiment data.
View Project →