Headline

AI Product Lead @ E42 | Agentic AI in production RAG, multi-agent orchestration, on-prem LLMs | Building “self-operating finance”

About

I build agentic AI that actually runs in production — powerful and practical, not just demos. I lead Neil at E42: an AI co-worker that runs the entire accounts-payable process. 96% extraction accuracy, template-free across any invoice layout (handwritten, multilingual, poorly scanned), N-way PO matching, and autonomous vendor-query resolution saving enterprises ~100 person-days a month. My path wasn't linear. I trained as a mechanical engineer, started out as design engineer, and taught myself into AI through a brutal year of 12-hour days. At E42 I went AI Engineer → Lead R&D → Product Lead — building the company's first RAG system, introducing the LLM stack that pivoted us to defensible on-prem deployment (great models, no supercomputer required), and architecting the foundation that became Neil. What I care about: multi-agent orchestration · RAG & retrieval · LLM systems in production · AI product strategy · enterprise finance automation. I stay close to the frontier agents, MCP, open models but I judge everything by one question: does it ship, and does it hold up with real customers? I'm building toward "self-operating finance," and writing about what it actually takes to put agentic AI in front of enterprises. If you're working on hard AI problems, AP/finance automation, or just believe in building AI that genuinely helps people let's connect.

Location

Mumbai

AI stack

claude code

hermes agent

langgraph

mcp

openai

huggingface

Work experience

Product Lead at E42.ai, May 2025 - Present · 1 yr 4 mos

Lead R&D Engineer at E42.ai, Aug 2022 - Jun 2025 · 2 yrs 11 mos, Pune, Maharashtra, India

AI/NLP Engineer at E42.ai, Jan 2022 - Jul 2022 · 7 mos, Pune, Maharashtra, India

AI and Machine Learning Intern at Lannet Technologies Pvt Ltd, Aug 2021 - Dec 2021 · 5 mos

Machine Learning Intern at Analysed.in, Mar 2021 - Jun 2021 · 4 mos, Mumbai, Maharashtra, India, 1. Built an AI and selenium based bot which scrape profile based on the job description uploaded by user and chat with them if they satisfy criteria. 2. Predicting the engagement rate using ML models of a post uploaded by user and testing user digital marketing skills. 3. Testing user SEO skills using NLP and giving score based on article uploaded by user and also checking for plagiarism.

Data Science Intern at The Sparks Foundation, Mar 2021 - Apr 2021 · 2 mos, Mumbai, Maharashtra, India, 1. Performed EDA to find the weak area of a business, to find out the deriving factors which is contribute to profit and which factor is contributing on loss. Recommend business solution through analysis 2. Building machine learning models to solve business related probelms

Design Engineer at DOUSE FIRE SYSTEMS PRIVATE LIMITED, Jan 2019 - Nov 2020 · 1 yr 11 mos

Education

Postgraduate Degree, Artificial Intelligence and Machine Learning, International Institute of Information Technology Bangalore

Bachelor of Engineering (B.E.), Mechanical Engineering, university of pune

Public links

github: https://github.com/rupeshpoojary9