- RAG-Enhanced-Chatbot-with-LoRA-Fine-Tuning — A powerful chatbot using Retrieval Augmented Generation, LoRA fine-tuning, and OpenAI embeddings.
- Architectures of Large Language Models
- Building Neural Networks from scratch using PyTorch + NumPy
- Developing Vector Databases and RAG Pipelines
| Project | Tech | Description |
|---|---|---|
| RAG Chatbot + LoRA | RAG, LoRA, OpenAI, React |
Conversational bot using Retrieval-Augmented Generation and LoRA fine-tuning |
| CourseCove (Spring boot + React) | Spring-Boot, React, Search Algorithms, Performance Optimization |
A semantic course search and recommendation system for online learning platforms, featuring intelligent query analysis, ranking, and autocomplete using Java and custom data structures |
| DeepSort - Realtime Object Tracking | Python, PyTorch, ResNet-50, YoloV4, OpenCV, CUDA, Kalman-Filters, Scikit-learn |
Implementation of real time object tracking in presence of occlusion - DeepSort |
| Extractive Text Summarizer | Python, NLP, PyTorch, SVM, Scikit-learn, NumPy, sciPy , NLTK |
Extractive Summarizer using SVM, trained on BBC news article dataset |
- Generative AI (Chatbots, LLMs, LoRA, RAG)
- Neural Network Architectures
- Software Engineering & Cloud Microservices
- MLOps and Scalable Model Deployment
- Open Source Collaboration
- Email: bhavsa85@uwindsor.ca | kathanbhavsar234@gmail.com
- Kaggle | LeetCode | GFG
I've built neural networks, fine-tuned LLMs, deployed ML microservices, and still enjoy solving a good LeetCode puzzle at midnight 🧩.