Hello, I'm
Tanuj Sansare
I build production-grade AI systems — from LLM agents and RAG pipelines to vision transformers. Passionate about turning cutting-edge ML research into scalable solutions that ship to millions.
Who I Am
I'm an AI Engineer specializing in LLMs, Vision Transformers, and RAG pipelines. At Bajaj Finserv, I've built systems generating 9,000+ personalized banners daily and AI agents automating enterprise workflows. I turn cutting-edge research into production-ready ML systems.
Featured Projects
A selection of my recent work
Technologies I Work With
My Experience
Professional background and education
AI Engineer — Content AI Pod
Bajaj Finserv
Pune, India
Leading AI-driven automation initiatives including Gen AI Bannerization (9,000+ daily assets) and Release Agent (autonomous DevOps). Architecting production-scale ML systems with Vision Transformers, LangGraph, and distributed processing.
- Built Gen AI Bannerization platform: 6 hours → 1 minute banner creation, 70% manual effort reduction
- Achieved 98% QC accuracy with fine-tuned Vision Transformer model for design defect detection
- Architected RabbitMQ + Celery distributed system handling 100+ simultaneous AI tasks
- Developed Release Agent using LangGraph/AutoGen, improving developer velocity by ~15%
AI Engineer — Content AI Pod
Bajaj Finserv
Pune, India
Leading AI-driven automation initiatives including Gen AI Bannerization (9,000+ daily assets) and Release Agent (autonomous DevOps). Architecting production-scale ML systems with Vision Transformers, LangGraph, and distributed processing.
- Built Gen AI Bannerization platform: 6 hours → 1 minute banner creation, 70% manual effort reduction
- Achieved 98% QC accuracy with fine-tuned Vision Transformer model for design defect detection
- Architected RabbitMQ + Celery distributed system handling 100+ simultaneous AI tasks
- Developed Release Agent using LangGraph/AutoGen, improving developer velocity by ~15%
Software Development Intern
Bajaj Finserv
Pune, India
Built enterprise-scale full-stack applications including KPI Portal (2M+ concurrent users) and Core Web Vitals optimization achieving 68% performance improvement.
- Developed KPI Portal with real-time analytics dashboard supporting 2M+ concurrent users
- Implemented AI-driven insights with secure data isolation and audit trails
- Achieved 68% improvement in Core Web Vitals (LCP, FID, CLS) through optimization
- Built WebAuthn passwordless authentication system for enhanced security
Software Development Intern
Bajaj Finserv
Pune, India
Built enterprise-scale full-stack applications including KPI Portal (2M+ concurrent users) and Core Web Vitals optimization achieving 68% performance improvement.
- Developed KPI Portal with real-time analytics dashboard supporting 2M+ concurrent users
- Implemented AI-driven insights with secure data isolation and audit trails
- Achieved 68% improvement in Core Web Vitals (LCP, FID, CLS) through optimization
- Built WebAuthn passwordless authentication system for enhanced security
Research Intern — Samsung PRISM
Samsung Research
Pune, India (Remote)
Conducted ML research on chat screenshot understanding, developing Sentimeter AI — a multi-stage OCR-NLP pipeline for enterprise sentiment intelligence.
- Built OCR-NLP pipeline achieving 0.92+ F1-score across WhatsApp, Slack, Samsung Messages
- Fine-tuned RoBERTa transformer for chat-specific sentiment classification
- Developed novel speaker diarization algorithm with 97.1% attribution accuracy
- Implemented privacy-first PII redaction module for GDPR/CCPA compliance
Research Intern — Samsung PRISM
Samsung Research
Pune, India (Remote)
Conducted ML research on chat screenshot understanding, developing Sentimeter AI — a multi-stage OCR-NLP pipeline for enterprise sentiment intelligence.
- Built OCR-NLP pipeline achieving 0.92+ F1-score across WhatsApp, Slack, Samsung Messages
- Fine-tuned RoBERTa transformer for chat-specific sentiment classification
- Developed novel speaker diarization algorithm with 97.1% attribution accuracy
- Implemented privacy-first PII redaction module for GDPR/CCPA compliance
AI/ML Intern
American Axle Manufacturing
Pune, India
Built Chat With Your Files — a production-grade RAG application enabling natural language queries over enterprise documents, completing full research-to-deployment in 8 weeks.
- Developed RAG pipeline using LangChain, FAISS, and Cohere API
- Reduced information retrieval time from 45 minutes to under 2 minutes
- Implemented multi-format document support (PDF, CSV, DOCX) with semantic chunking
- Delivered production-ready Streamlit interface with conversational memory
AI/ML Intern
American Axle Manufacturing
Pune, India
Built Chat With Your Files — a production-grade RAG application enabling natural language queries over enterprise documents, completing full research-to-deployment in 8 weeks.
- Developed RAG pipeline using LangChain, FAISS, and Cohere API
- Reduced information retrieval time from 45 minutes to under 2 minutes
- Implemented multi-format document support (PDF, CSV, DOCX) with semantic chunking
- Delivered production-ready Streamlit interface with conversational memory
B.Tech Computer Engineering
College of Engineering Pune (CoEP)
Pune, India
Graduated with focus on AI/ML, software engineering, and data structures. Participated in Samsung PRISM research program.
- Samsung PRISM Research Program participant
- Focus areas: Machine Learning, NLP, Computer Vision
- Capstone: Enterprise RAG system for document intelligence
B.Tech Computer Engineering
College of Engineering Pune (CoEP)
Pune, India
Graduated with focus on AI/ML, software engineering, and data structures. Participated in Samsung PRISM research program.
- Samsung PRISM Research Program participant
- Focus areas: Machine Learning, NLP, Computer Vision
- Capstone: Enterprise RAG system for document intelligence
Let's Work Together
Have a project in mind? Let's talk