Headline

Founding AI & Backend Engineer · Building GuppShupp

About

Long story short: I’m that guy who takes boring, 3-hour manual work and turns it into “click, query, ho gaya” using GenAI, RAG, and a lot of carefully managed chaos. I work as an AI/ML Engineer, mostly living in the world of RAG systems, LangGraph workflows, OCR pipelines, and LLM fine-tuning. If the problem sounds like “humans are doing this on Excel + WhatsApp + pain”, that’s usually where I enter. Recent scenes: - Built Jeevani, a production focused RAG platform where gaushalas and agri teams ask biomass questions in plain language and get SQL-backed answers instead of 3+ hour report marathons. 1000+ facilities, LangGraph orchestration, hybrid search, the full jugaad-to-production arc. - Shipped FormExtract AI, a 6-stage Azure Document Intelligence + Gemini pipeline that reads messy forms, extracts key-values, lets humans quickly validate, and spits out clean Excel/JSON/PDF. Time saved? Let’s just say ops teams stopped hating forms that much. - Built Indic-Voice-AI, a multilingual voice companion that understands 18 Indic languages, remembers context, detects emotion, and replies with natural TTS — all in ~6–10 seconds end-to-end on FastAPI + pgvector + LangGraph. - Fine-tuned Qwen 2.5 Coder 7B with QLoRA on 10K+ samples, squeezed it into consumer-GPU budgets with 4-bit quantization, and still got big jumps in task accuracy. Basically, model ko gym bheja aur RAM bhi bacha li (Model bhi happy, wallet bhi). Tech stack wise, I’m most at home with Python, FastAPI, LangChain/LangGraph, RAG pipelines, vector DBs (FAISS/Chroma/pgvector), Hugging Face Transformers/PEFT, Whisper, Azure Document Intelligence, Gemini, PostgreSQL, Docker. I’ve applied this across internships and production projects at SSH Softtech and UptoSkills, not just side demos. On paper: BCA, 9.18/10 CGPA (Top 1%) from Christ Academy, plus 200+ hours of focused training in RAG, LLM fine-tuning, LangGraph, ML and DL. In reality: obsessive debugger, late-night experiment guy, and the friend people call when “LLM se kuch karna hai par pata nahi kya”. I’m looking for GenAI / AI Engineer roles where the expectation is: own the problem, design the data and architecture, build the RAG/LLM/voice/vision pieces, and ship something actual users rely on — not just another “hello world” chatbot. If that’s your scene too, DM open. Pata nhi kyu par LAGAV SA HO GAYA HEI LLMs ke SAATH 🫂 (maybe because it's a friend you can TRUST on 💯)

Location

Bengaluru

Work experience

Founding AI Engineer, Backend Engineer at GuppShupp - Life Companion, May 2026 - Present · 4 mos, Bengaluru

Full Stack AI Engineer at SSH SOFTTECH SOLUTION, Nov 2025 - Present · 10 mos, Bengaluru, • Spearheaded evolution of production RAG system from single-domain (gaushala) to multi-domain platform serving 1000+ facilities across Haryana and UP, scaling pan-India, expanding to crop analytics based on customer requirements while maintaining 95% query accuracy • Engineered intelligent query routing system using LangGraph multi-agent orchestration, enabling dynamic cross-database querying across gaushala and crop datasets, autonomously determining single-domain versus multi-domain execution paths based on semantic query intent • Architected NL-to-SQL conversational AI agent with Chroma vector search, translating natural language to contextually aware SQL across heterogeneous schemas, handling livestock biomass, crop yield analysis, and cross-domain agricultural insights in a unified interface • Automated end-to-end analytics pipeline with async FastAPI backend, reducing manual PDF and Excel processing from 3 hours to 5 minutes (97% time reduction) and cutting monthly reporting cycles by 70% through Streamlit dashboard deployment Tech Stack: Python | LangGraph | FastAPI | PostgreSQL | Chroma | Streamlit

AI Engineer Intern at GuppShupp - Life Companion, Feb 2026 - Apr 2026 · 3 mos, Bangalore Urban

Project Based AI Engineer (Contract) at GuppShupp - Life Companion, Dec 2025 - Jan 2026 · 2 mos, Bangalore Urban, Karnataka, India

AI Engineer Intern at UptoSkills, Nov 2025 - Feb 2026 · 4 mos, Bengaluru, Karnataka, India, • Architecting production RAG pipeline for medical Q&A serving 50+ healthcare queries daily (85% retrieval accuracy) • Fine-tuning 7B LLMs using QLoRA on 10K+ samples (40% memory reduction, 18% accuracy gain) • Optimized inference latency by 35% via batch processing + 4-bit quantization for real-time clinical queries Tech: Python | PyTorch | LangChain | FAISS/Chroma | Hugging Face | QLoRA

Generative AI Trainee at Ei Systems, Oct 2025 - Nov 2025 · 2 mos, Remote

Data Science Intern at Dyashin Technosoft, Feb 2025 - Mar 2025 · 2 mos, Bangalore Urban, Karnataka, India, Data Scientist & Data Analyst – Dyashin Technosoft At Dyashin Technosoft, I drove end-to-end AI/ML initiatives—transforming raw data into strategic insights and fully deployed models. Engineered robust data pipelines with Pandas and NumPy, automating preprocessing of 10 k+ records and reducing manual effort by 40%. Designed, trained & fine-tuned classification/regression models using Scikit-Learn and TensorFlow, achieving 90%+ accuracy via systematic hyperparameter optimization. Implemented customer segmentation with K-Means clustering, enabling targeted campaigns that boosted ROI by 25%. Built interactive dashboards (Matplotlib, Plotly) to visualize key metrics, accelerating data-driven decision-making across marketing and product teams. Collaborated cross-functionally: led weekly sprint demos, documented workflows, and enforced Git/GitHub best practices for reproducible code. Demonstrated eagerness to learn by prototyping NLP/Transformer techniques and contributing to peer code reviews and knowledge-sharing sessions. This role honed my analytical rigor, model deployment expertise, and passion for leveraging AI to solve real-world business challenges.

Education

Bachelor of Computer Application, Degree, Christ Academy Institute for Advanced Studies

CEBA, 12th, Christ Academy junior college