Open to opportunities

Debanjan Chaudhuri

@debanjanchaudhuri

AI Architect and Senior Data Scientist specializing in Agentic AI, NLP, and computer vision—delivering procurement automation and cost savings.

Germany

Message

What I'm looking for

I’m looking for an AI-first team where I can build agentic NLP/vision systems end-to-end—semantic search, knowledge graphs, and procurement automation—shipping production-ready models that reduce cost and improve decision-making through measurable impact.

I’m an AI Architect and Senior Data Scientist with 16+ years of experience, including 10 years in Natural Language Processing (NLP) and Computer Vision. I bridge industry and academia, consistently building AI-driven solutions to tackle complex business and research problems—especially where LLMs, semantic search, and knowledge graphs can create measurable impact.

In my current role at Uniper, I lead AI-first initiatives within Procurement and developed “Procurement Buddy,” a domain-specific, LLM-based agent that enables FAQ resolution and SAP catalog purchases with semantic search, integrating APIs and knowledge sources to optimize the procurement process by 85%. I’ve also helped productionize ML for timed delivery prediction, implemented OCR-based invoice automation (with signature verification and handwritten text detection), and guided end-to-end agentic ML implementations to drive automation and efficiency gains that deliver significant cost savings—while working with open-source market data for cost optimization.

Experience

Work history, roles, and key accomplishments

Current

AI Architect

Current

Uniper

Jan 2025 - Present (1 year 5 months)

Led AI-first Procurement initiatives by developing Procurement Buddy, an LLM-based agent for FAQ resolution and SAP catalog purchases with semantic search, optimizing the procurement process by 85%. Built production-ready ML to predict purchase order delivery timing and implemented cost-optimization negotiation strategies using open market data (Destatis).

Agentic AI LLM Semantic Search SAP Ariba Cost Optimization API Integration Python

Senior Data Scientist

Easy2Parts

May 2023 - Dec 2024 (1 year 7 months)

Architected a knowledge graph-powered product pipeline using local LLMs (GLiNER) to extract entities and relations from engineering drawings, improving price prediction accuracy by 7%. Developed OCR-based drawing segmentation to enhance detection of GD&T and technical symbols, further improving price prediction.

Knowledge Graphs Entity Relation Extraction Computer Vision Price Prediction data preprocessing Python

Senior Data Scientist

RTL

Jul 2022 - Apr 2023 (9 months)

Spearheaded semantic search and knowledge graph solutions for RTL+ by building a music and podcast search engine with a 700 ms performance gain while scaling retrieval across 5B+ entities. Engineered a cost-efficient semantic search system over knowledge graphs to improve music and movie recommendation quality.

Semantic Search Knowledge Graphs Information Retrieval Recommendation Systems Entity Linking Performance Optimization Python

AI/Digitalization Expert

Uniper

Aug 2020 - Jun 2022 (1 year 10 months)

Developed PIA, an NLP-powered chatbot for procurement FAQs and SAP Ariba catalog search. Automated invoice processing using Azure OCR (signature verification and handwritten text detection) integrated with RPA workflows, and built a fast contract search tool using text-matching algorithms (RapidFuzz).

NLP SAP Ariba Azure OCR RPA Document Processing Python

NLP Consultant

Freelancer

Jan 2020 - Jan 2022 (2 years)

Built a question-answering system for personal documents using SOTA reading comprehension models, semantic search, and PII obfuscation. Delivered a NASA-contracted entity-extraction product from patents and news to enable investment trend analysis and stock price forecasting.

Reading Models Semantic Search Information Retrieval Forecasting

LP Researcher

Fraunhofer Iais

Jul 2017 - Jul 2020 (3 years)

Pioneered entity and relation linking algorithms for natural language questions aligned with PhD research. Collaborated with Volkswagen to design an in-car conversational assistant using knowledge graphs for entity-aware Q&A about city points of interest.

Entity Relation Linking Knowledge Graphs Conversational AI Natural Language Processing Dialogue Systems Research Engineering Python

Machine Learning Researcher

Wipro

Jan 2016 - Nov 2016 (10 months)

Researched argument mining from heterogeneous sources (web/forums/social media) using CNNs/LSTMs and worked on fake news detection for an academic COLING paper.

CNN Natural Language Processing Machine Learning Research Deep Learning

Software Developer

Atos

Jan 2009 - Jan 2015 (6 years)

Engineered CRM backend systems for E-Plus Germany using PL/SQL and bash scripting, and implemented early AI to automate support ticket classification. Contributed to software engineering efforts that accelerated issue resolution with text-based analytics.

Pl SQL Bash Scripting SQL Support Ticket Classification Backend Engineering Software Development

Education

Degrees, certifications, and relevant coursework

University of Bonn

Doctor of Philosophy (Ph.D.), Natural Language Processing

2017 - 2022

Completed a Ph.D. focused on enriching text-based human-machine interactions with additional world knowledge, advancing research in NLP and related generation/grounding approaches.

BITS Pilani

Master of Technology (M.Tech), Software Systems (Data Analytics)

2014 - 2016

Earned an M.Tech in Software Systems with a Data Analytics specialization, building expertise in analytics-focused engineering and software systems.

Kalyani Government Engineering College

Bachelor of Technology (B.Tech), Electronics and Telecommunication Engineering

2005 - 2009

Completed a B.Tech in Electronics and Telecommunication Engineering, with foundation coursework supporting subsequent work in computer vision and OCR.