Dimitris Gkoumas, PhD
Staff ML Engineer & AI Research Scientist · Reading, UK
About
I build and research LLM systems: agentic AI, retrieval-augmented generation, LLM post-training and evaluation, and multimodal learning. At Method (Hitachi Digital) I lead end-to-end AI solution development, building AI systems that capture organisational knowledge and turn it into actionable AI capabilities.
Before that I was a Research Scientist at Huawei Noah's Ark Lab, working on LLM reasoning with reinforcement learning and agentic workflows, and a Postdoctoral Research Associate at Queen Mary University of London, working on longitudinal NLP. I hold a PhD in Artificial Intelligence from The Open University (Marie Curie scholarship) on quantum-inspired multimodal representation learning and have published at ACL, EMNLP, AAAI and IJCAI.
AI Under the Hood Study notes on ML foundations, why models fail to generalise, how they are optimised, how generative models and transformers work, how training scales across GPUs, how it all comes together in LLMs, how to deploy them, and how agents use tools. Read →Experience
- 2025 – nowStaff ML Engineer, Method – Hitachi Digital, LondonEnterprise AI solutions; multi-agent design systems; production AI and MLOps.
- 2024 – 2025Research Scientist in NLP, Huawei Noah's Ark Lab, LondonLLM mathematical reasoning: reward modelling, RL and execution feedback; agentic code repair.
- 2021 – 2024Postdoctoral Research Associate, Queen Mary University of LondonLongitudinal NLP; RAG, parameter-efficient fine-tuning, RLHF, efficient inference, hallucination evaluation.
- 2021 – 2023Applied ML Engineer, Huawei Ireland Research Centre, DublinLarge-scale semantic search; distributed LLM training; ML infrastructure.
- 2017 – 2021Early Stage Researcher, The Open UniversityQuantum-inspired neural models for multimodal learning.
Education
- 2017 – 2021PhD in Artificial Intelligence, The Open University, UKQuantum-inspired multimodal representation learning.
- 2012 – 2014MSc in Machine Learning, Aristotle University of Thessaloniki
- 2005 – 2009BSc in Computer Science, Technical University of Crete
Selected publications
- Less for More: Enhanced Feedback-aligned Mixed LLMs for Molecule Caption Generation and Fine-Grained NLI Evaluation ACL 2025 (oral)
- Reformulating NLP Tasks to Capture Longitudinal Manifestation of Language Disorders in People with Dementia EMNLP 2023 (oral)
- A Digital Language Coherence Marker for Monitoring Dementia EMNLP 2023
- A Longitudinal Multi-modal Dataset for Dementia Monitoring and Diagnosis Language Resources and Evaluation, 2024
- An Entanglement-driven Fusion Neural Network for Video Sentiment Analysis IJCAI 2021
- A Quantum Cognitively Motivated Decision Fusion Framework for Video Sentiment Analysis AAAI 2021
Full list on Google Scholar.