Open to opportunities

Stamatios Batelis

@stamatiosbatelis

Lead Data Scientist specializing in LLMs, NLP, and GenAI—building production ML that turns research into business value.

Greece

Message

What I'm looking for

I’m looking to lead GenAI/NLP work that ships production LLM applications—owned end to end—with measurable business impact (risk, cost, and accuracy), strong cross-functional collaboration, and room to mentor teams and scale capability.

I’m a Lead Data Scientist specializing in Large Language Models (LLMs), NLP, and Generative AI (GenAI), focused on architecting and deploying production-grade ML solutions. In fintech and regulated environments, I’ve led AI Core roadmaps for GenAI integration, built and optimized customer-facing chatbots, and strengthened fraud detection models to reduce financial losses. I consistently bridge complex research and commercial value—turning advanced models into measurable outcomes like cost savings and improved extraction accuracy.

Most recently, I led LLM-based deployments for text summarization and interactive chatbots, and delivered automated document analysis such as an automated blacklining workflow for high-risk discrepancy detection. Earlier roles include re-architecting an RFQ engine by replacing legacy SpaCy models with a fine-tuned T5-family LLM via LoRA (plus custom data augmentation), and building enforcement entity extraction with SpaCy and Streamlit alongside proceeds-of-crime detection using YOLOv5 and Faster R-CNN. I also lead through mentorship—scaling teams via technical training and recruitment—and I bring a researcher’s discipline from my PhD background in improving hydrological processes with spatial-temporal big data.

Experience

Work history, roles, and key accomplishments

Current

Lead Data Scientist

Current

Plum Fintech

Nov 2024 - Present (1 year 7 months)

Led the AI Core team, defining the GenAI roadmap across the product suite and directing customer-facing chatbot development, optimization, and evaluation. Enhanced critical fraud-detection ML models to significantly reduce financial losses while scaling the data science function through revised hiring and planning frameworks.

Generative AI NLP Chatbot Development Model Evaluation Fraud Detection Roadmapping Team Leadership

Lead Data Scientist

HSBC

Nov 2023 - Sep 2024 (10 months)

Led deployment of LLM-based applications for text summarization and interactive customer chatbots within Market & Security Services. Delivered automated blacklining for high-risk legal document discrepancies and built a RAG market guide assistant (HuggingFace, LangChain) saving £500K per year in operational costs, while mentoring and scaling the team.

LLM Summarization Conversational AI RAG LLMs (LangChain Document Automation Risk Focused ML Mentoring

Senior Data Scientist

HSBC

Sep 2022 - Oct 2023 (1 year 1 month)

Modernized the RFQ engine by replacing legacy SpaCy models with a fine-tuned T5-family LLM using LoRA, improving extraction accuracy by 5%. Built custom data augmentation for an in-house NLP platform and collaborated with global economists and ML experts using OpenAI and LangChain to research and implement new algorithms.

SpaCy OpenAI LLMs (LangChain NLP Platform Engineering Cross Functional Collaboration

Senior Associate Data Scientist

Financial Conduct Authority (Fca)

Aug 2020 - Sep 2022 (2 years 1 month)

Built and deployed an automated entity extraction tool using SpaCy and Streamlit to support enforcement investigations. Developed proceeds-of-crime identification using YOLOv5 and Faster R-CNN, led graph analysis (NetworkX, Pyvis) to uncover financial crime patterns, and managed/mentored a team of two junior data scientists with organization-wide technical training as SME.

SpaCy Streamlit YOLOv5 Proceeds Of Crime Detection Graph Networkx Technical Leadership YOLO V5

Data Scientist

British Transport Police

Sep 2019 - Jul 2020 (10 months)

Improved predictive accuracy for new railway line patronage by developing and deploying a custom regression-based forecasting model. Architected and tested an image recognition system for automated visual data processing using TensorFlow with VGG and ResNet frameworks, and ran ML seminars to upskill departmental staff.

Demand Forecasting Regression Modeling Computer Vision TensorFlow ResNet Image Recognition Machine Learning Technical Training predictive modeling

Teaching Assistant/PhD Researcher

School of Civil Engineering, University of Bristol

Sep 2016 - Jun 2019 (2 years 9 months)

Supervised multi-disciplinary student teams to deliver complex engineering projects and built Monte Carlo simulation frameworks to model and incorporate stochastic uncertainty into physical experiments. Provided instruction and mentorship on Python/Matlab and engineering principles, and authored peer-reviewed research papers while presenting findings at international conferences.

Project supervision Uncertainty Quantification Monte Carlo Simulations Python Matlab Technical Mentorship Research Writing

Education

Degrees, certifications, and relevant coursework

University of Bristol

Doctor of Philosophy (PhD), Civil Engineering

2015 - 2020

PhD in Civil Engineering (WISE CDT) focused on improving the hydrological processes of the UK land surface model JULES, analyzing spatial-temporal big data to develop and validate hydrological models.