Sanjit Kumar
AI/ML Engineer & Generative AI Specialist

AI/ML Engineer with hands-on project experience in Generative AI, LLM applications, RAG, and multi-agent systems. Builds AI applications using Python, FastAPI, LangChain, LangGraph, Hugging Face, and FAISS. Experienced in semantic retrieval, conversational memory, API integration, and model fine-tuning.

Generative AI & LLMs RAG Systems Multi-Agent Systems Fine-Tuning (LoRA / QLoRA) LangChain & LangGraph
Available for Freelance & Full-time Remote • Based in India
Explore Projects Contact

Tech Stack

Full toolkit — from data pipelines and model training to deployment and production APIs.

AI / ML / GenAI

Machine Learning NLP Transformers LangChain LangGraph RAG AI Agents FAISS Fine-Tuning LoRA / QLoRA LangSmith Prompt Engineering

Libraries & Frameworks

Scikit-learn PyTorch Hugging Face Pandas NumPy XGBoost LightGBM Matplotlib Plotly

Backend & APIs

Python FastAPI Django REST APIs SQLite JWT Auth SQL Streamlit

MLOps & DevOps

MLflow DVC Docker Git & GitHub AWS (Basic) Railway Render

About

AI/ML Engineer with hands-on project experience in Generative AI, LLM applications, RAG, and multi-agent systems. Built AI applications using Python, FastAPI, LangChain, LangGraph, Hugging Face, and FAISS. Experienced in semantic retrieval, conversational memory, API integration, LLM application development, and model fine-tuning.

Also the founder of csestudy247.com, a production-grade educational platform with a published Android app (100+ downloads on Google Play).

View Skills Let's Collaborate

Professional Highlights

  • Built multi-agent travel planner with LangGraph, Groq Llama-3.3-70B & FastAPI backend
  • INDGPT — full-stack conversational AI with JWT auth, SQLite persistence & real-time streaming
  • Semantic job matching with BERT + FAISS on 17,000+ postings — 85% accuracy
  • AI/ML skills: Fine-Tuning, LoRA, QLoRA, Prompt Engineering, LangSmith, Hugging Face
  • Founder of csestudy247.com — WordPress platform + Android app (100+ downloads)
  • B.Tech CSE CGPA 8.24 | Diploma CSE CGPA 8.68 — Nalanda College of Engineering
  • Certified in Data Science, Machine Learning & Generative AI — CampusX

Skills

A categorized view of my technical expertise — from data science and ML to Generative AI and MLOps.

🐍

Languages

Python SQL
📊

Data Science

EDA Feature Engineering Feature Selection Data Visualization Statistical Analysis Model Evaluation Cross Validation Hyperparameter Tuning
🤖

Machine Learning

Regression Classification Clustering Recommendation Systems Random Forest XGBoost LightGBM Stacking Regressor
✨

Generative AI

LangChain LangGraph RAG AI Agents Prompt Engineering Fine-Tuning LoRA QLoRA LangSmith Google Gemini Hugging Face Vector Databases
🧠

Deep Learning & NLP

PyTorch Neural Networks Transformers BERT Sentence Transformers TF-IDF Hugging Face
📦

Libraries

Pandas NumPy Scikit-learn Matplotlib Plotly FAISS XGBoost LightGBM
⚙️

MLOps

MLflow DVC Docker Pipeline Orchestration Experiment Tracking
🔧

Backend & Databases

FastAPI Django REST APIs SQLite JWT Git & GitHub VS Code Jupyter Notebook
☁️

Cloud & Deployment

AWS (Basic) Railway Render Hostinger Google Play Store

Projects

End-to-end AI and ML builds — from RAG systems and AI agents to production ML pipelines. All linked to GitHub.

Multi-Agent • LangGraph • FastAPI • 2026

AI Multi-Agent Travel Planner

A multi-agent travel planning system with specialized agents for flight search, hotel search, itinerary generation, and final trip planning — powered by Groq Llama-3.3-70B.

  • Specialized agents orchestrated with LangGraph persistent state
  • Integrated Groq Llama-3.3-70B, AviationStack & Tavily AI
  • FastAPI backend with real-time agent coordination
Python LangGraph FastAPI Groq Tavily AI
Multi-Agent System GitHub ↗

GenAI • FastAPI • LangGraph • 2026

INDGPT — Conversational AI App

Full-stack conversational AI application with persistent chat memory, multi-session conversations, JWT authentication, and real-time response streaming.

  • Persistent chat memory with SQLite and multi-session support
  • JWT authentication with guest workflows
  • Deployed on Render with real-time streaming via FastAPI
FastAPI LangGraph Gemini SQLite JWT
Live on Render GitHub ↗

AI • NLP • Semantic Retrieval • 2025

AI-Powered Recruitment System

Semantic job recommendation using Sentence-BERT embeddings and FAISS similarity search — processed over 17,000 job postings for intelligent candidate-job matching.

  • BERT-based semantic representations of candidate profiles & job descriptions
  • FAISS similarity search for scalable embedding-based retrieval
  • Processed 17,000+ job postings with 85% matching accuracy
Python Django BERT FAISS Sentence Transformers
85% Accuracy GitHub ↗

GenAI • RAG • Healthcare • 2026

Medical AI Assistant

RAG-based healthcare information assistant using FAISS vector search and LLM-powered response generation for grounded, context-aware answers.

  • Retrieval-Augmented Generation with FAISS for accurate medical responses
  • Conversational question answering by retrieving relevant medical context
  • LangChain pipeline with Google Gemini for response generation
Python LangChain RAG FAISS Gemini
Healthcare AI GitHub ↗

ML • Prediction • Real Estate

Prophe AI — Property Price Prediction

AI-powered real estate platform predicting property prices in Gurugram with interactive analytics and cutting-edge ML.

  • Supervised ML model for real-time property price prediction
  • Data preprocessing, feature engineering, and model evaluation
  • Deployed as a production-ready web application
Python Django Machine Learning Scikit-learn
Production-ready GitHub ↗

MLOps • ML Pipeline • Docker

Swiggy Delivery Time Prediction

End-to-end ML pipeline predicting food delivery time using traffic, weather, rider, and delivery features — fully containerized.

  • Stacking Regressor: Random Forest + LightGBM + Linear Regression
  • DVC for pipeline orchestration, MLflow for experiment tracking
  • Dockerized for consistent deployment
Python Scikit-learn MLflow DVC Docker
Full MLOps Stack GitHub ↗

NLP • Sentiment • MLOps

YouTube Comment Sentiment Analysis

End-to-end NLP pipeline for YouTube comment sentiment classification using TF-IDF, LightGBM, MLflow, and DVC.

  • TF-IDF feature extraction + LightGBM classification
  • MLflow for experiment tracking and model versioning
  • DVC for data & pipeline management
Python NLP LightGBM MLflow DVC
NLP Pipeline GitHub ↗

AI Agent • LangGraph • Data Engineering

AI Agent for Data Cleaning

AI-powered automated data cleaning system using LangGraph and LLMs for intelligent preprocessing of CSV, Excel, PostgreSQL, and REST API data.

  • LangGraph + Gemini/Ollama for automated preprocessing decisions
  • Handles missing values, duplicates, type correction, and validation
  • FastAPI backend + Streamlit UI for interactive usage
Python LangGraph FastAPI Streamlit Gemini
Agentic AI GitHub ↗

Platform • WordPress • SEO

csestudy247.com — Educational Platform

Production-grade educational platform built and maintained from 2022. Custom plugins, online assessments, and ML/web app deployments.

  • Custom plugins for online assessments and test management
  • SEO and analytics strategies for platform growth
  • Deployed ML and web apps on Railway, Render, and Hostinger
WordPress PHP MySQL SEO
Founder • 2022–Present Visit ↗

Contact

Looking to hire a Data Scientist or AI/ML Engineer? Want to collaborate on a project or build something impactful? Let's connect — I'm open to full-time roles, freelance work, and research collaborations.

📧 Email Me LinkedIn
+91-9304533853 Hindi & English

Links

GitHub
Public projects & contributions LinkedIn
Connect professionally Hugging Face
Models & datasets Website (csestudy247.com)
Educational platform Email
sanjitchaudhary506@gmail.com
Scroll to Top