Anish Rane

LLM-AI-Analytic-Project

A LangChain + OpenAI-powered agent that enables natural language analysis of sales data with built-in forecasting, anomaly detection, profitability, and reporting tools.

Jupyter Notebook

Gesture-Recognition-using-3D-CNN

This repository contains experiments with spatiotemporal deep learning models for sequence-based tasks. Various architectures, including Conv3D, ConvLSTM, and hybrid models with GRU/LSTM, were tested with different regularization techniques. The best model achieved 78% validation accuracy.

Jupyter Notebook

Insurance-HelpMateAI

An end-to-end Retrieval-Augmented Generation (RAG) system built to query complex insurance documents. Combines OpenAI embeddings, ChromaDB for vector search, a cross-encoder for reranking, and GPT-3.5 for grounded, context-aware answers.

Jupyter Notebook

Cylindrical-Filter-Simulator

A high-fidelity Python simulator for cylindrical air filters with fluid dynamics, dust loading, material database, and exportable research-grade plots and reports.

Jupyter Notebook • 1 star

RAG-Fine-Tuning

Production-minded RAG prototype: embedding ingestion, vector search, LoRA/PEFT fine-tuning and a Node.js orchestrator API serving context-aware chat replies.

Python

Skin-Cancer-Detection-using-CNN

This repository contains a deep learning model built on a Convolutional Neural Network (CNN) architecture for Skin Cancer Detection. The model is trained to classify skin lesions into different categories based on medical image data.

Jupyter Notebook