03 — Projects
Resuvanta AI Resume Screener
NLP-based candidate screening system using TF-IDF vectorization and classification models to rank candidates against job descriptions. Automated feature extraction and scoring eliminate manual shortlisting effort.
RAG based AI Assistant
A Retrieval-Augmented Generation (RAG) system that enables intelligent question answering over video content by leveraging audio transcription, semantic embeddings, and LLM-based response generation.
ClipzeeAI — Reel Generator
Full-stack automated reel generation system that converts text and images into audio-narrated video content. Flask-powered backend orchestrates multi-stage media processing — text-to-audio synthesis, image sequencing, and final rendering which helps in decreasing average video production time by 80%.
Depression Detection using ML
Supervised machine learning model developed during our Academic Project which classifyies depression indicators from text datasets (1K--10K samples). Compared SVM (63.5% accuracy, F1: 38.8%) and CNN on 1K–10K text samples; the CNN achieved 93.1% accuracy and F1: 92.7% -- a 29.6-point gain over baseline.