🔭 I'm currently pursuing a Master's in Computer Science at NYU Tandon School of Engineering, specializing in neuroinformatics along with computer vision, machine learning, and Large Language Models (LLMs).
- GPA: 4.0 (First Semester)
- Courses: Design and Analysis of Algorithms, Machine Learning, and Software Engineering
🎓 B.E (Hons.) in Electrical Engineering from Jadavpur University
- GPA: 8.78/10.0
🧠 Previously, I worked as a Computer Vision Engineer at BigVision.ai (June 2022 - June 2024), where I:
- Developed people and vehicle detection systems using YOLOv5n with state-of-the-art F1-scores on infrared video datasets
- Enhanced YOLOv5 models with Alarm-based and Kalman filter-based tracking for precise object tracking
- Optimized video encoding/decoding pipelines using OpenVINO, FFmpeg, and OpenCV
- Cut OpenAI API costs 20x by shifting from GPT-4 to GPT-3.5 with advanced prompt engineering
🌟 Also served as a Mitacs Globalink Research Intern at ÉTS Montreal (June 2021 - Sept 2021):
- Developed 360StereoNet, the first model to generate depth maps from 360-degree stereoscopic images
🌱 My academic and professional journey is committed to exploring and harnessing the transformative capabilities of AI technologies to develop innovative solutions for both digital and physical environments.
💬 I'm actively seeking collaborative opportunities in research and projects that advance the frontiers of artificial intelligence, with a focus on real-world applications and theoretical advancements.
- 🖼️ Computer Vision
- 🤖 Machine Learning
- 🧠 Large Language Models
- 🔍 Human Pose Estimation
- 📊 3D Vision & Reconstruction
- Conducted 50 user interviews and surveys to validate product vision and define requirements for an MVP aimed at democratizing personal styling
- Built a scalable backend using Supabase and FastAPI, and developed the frontend with HTML, CSS, and JavaScript
- Collaborated with 3 professional stylists to create digital avatars using LangGraph to build AI Agents which are clones of actual professional stylists
- Scraped data from popular fashion e-commerce websites (around 8k products), converted their descriptions into embeddings, and stored them in Pinecone for an LLM-based fashion recommendation system
- Deployed on AWS
- 🔗 Website Link
- Built an AI-powered university at the Cornell AI Hackathon, featuring an AI professor, AI teaching assistant, and AI auditor working together to evaluate students and personalize learning
- The AI professor teaches using uploaded slides/PDFs, the AI teaching assistant generates quizzes, and the AI auditor evaluates student progress, adjusting teaching style accordingly
- Implemented the backend with FastAPI, LangChain, and Python, and built the frontend with React
- 🔗 Curio Repository
An Optimized Fuzzy Ensemble of Convolutional Neural Networks for Detecting Tuberculosis from Chest X-ray Images (January 2022)
- Developed a model for TB detection from Chest X-rays using a type-1 Sugeno fuzzy integral ensemble method
- Integrated outputs from DenseNet121, VGG19, and ResNet50
- Surpassed existing techniques on a public TB dataset
- Published in Applied Soft Computing, Elsevier
- 45+ Google Scholar citations
💼 Interested in collaborating on research? Feel free to reach out: [email protected]