Portfolio & Selected Works

Featured Engineering Projects

Explore production AI recruiting workflows, advanced RAG systems, semantic search, agentic AI experiments, and scalable full-stack products built around real user needs.

Explore Projects View on GitHub

Flagship Work Experience

Shuvel Digitech

AI & FULL STACK RECRUITING

GiFTEM – AI-Powered Recruiting Platform

Enterprise AI platform helping recruiters source, analyze, shortlist, and engage candidates using semantic discovery, match analysis, skill-gap insights, and AI outreach.

Next.js

TypeScript

Node.js

PostgreSQL

4 Modes

Semantic Search

25–40%

Speed Boost

AI Voice

Outreach Auto

Curated Portfolio

Selected Work

Production-ready AI and full-stack systems that solve real problems and deliver measurable impact.

Advanced RAG Pipeline for Multi-PDF Question Answering project preview

AI Projects

Production Ready

Advanced RAG Pipeline for Multi-PDF Question Answering

A multi-document RAG pipeline built with LangChain, SentenceTransformers, ChromaDB, and Groq Llama 3.1. It recursively loads PDFs, preserves source metadata, splits content into overlapping chunks, generates local embeddings, performs filtered semantic retrieval, and produces context-grounded answers with page references and supporting evidence.

Python

LangChain

ChromaDB

SentenceTransformers

all-MiniLM-L6-v2

Hugging Face

Groq

Llama 3.1

+8 more

View Case Study Code

AskMyDocs — Multi-Document RAG Assistant project preview

AI Projects

Production Ready

AskMyDocs — Multi-Document RAG Assistant

Multi-document question-answering application using LangChain document loaders, recursive text chunking, SentenceTransformer embeddings, FAISS semantic retrieval, Groq-hosted Llama models, and an interactive Streamlit interface.

Python

Streamlit

LangChain

FAISS

SentenceTransformers

Hugging Face

Groq

Llama 3.1

+5 more

View Case Study Code

Vectorless RAG Pipeline with PageIndex project preview

AI Projects

Production Ready

Vectorless RAG Pipeline with PageIndex

A structure-aware RAG pipeline that converts PDF documents into hierarchical trees, identifies relevant sections through LLM-based node selection, and generates grounded answers with section titles and page references using Groq Llama 3.1 and LangChain.

Python

LangChain

PageIndex

Groq

Llama 3.1

Vectorless RAG

Hierarchical Retrieval

LLM Tree Search

+2 more

View Case Study Code

Let's build something impactful

I'm open to AI engineering, full-stack AI, and applied GenAI opportunities and collaborations.

Get In Touch Explore Services