Demetris Paschalides
Lead ML & NLP Engineer at Laboratory for Internet Computing
Nicosia, Cyprus
Summary
👤
Senior🎓
Top SchoolDemetris Paschalides is a Lead ML & NLP Engineer with 11 years of experience and over nine years focused on end-to-end NLP pipelines, transformer model development, LLM fine-tuning, and production ML systems using PyTorch, HuggingFace, Kafka, Docker, and AWS. He combines research rigor from a PhD in Computer Science with hands-on engineering, deploying LLaMA/Mistral/Qwen-based systems, INT8-optimized ONNX inference, and serverless architectures for real-time streaming applications. His work spans social media, misinformation detection, cybersecurity, energy, and labor markets, often enriching models with GNNs and knowledge graphs via PyTorch Geometric and RAG. A collaborative R&D lead, he has coordinated projects with 50+ partners, supervised multiple theses, and reviewed for top conferences, translating academic insight into operational ML products. Based in Nicosia, Cyprus, he’s equally comfortable tuning models for cloud and on-device inference (TensorFlow.js), reflecting a practical bent for moving novel research into production.
11 years of coding experience
9 years of employment as a software developer
Doctor of Philosophy - PhD, Computer Science, Doctor of Philosophy - PhD, Computer Science at University of CyprusLyceum Deploma, Computer Science, Lyceum Deploma, Computer Science at Palouriotissa Lyceum
Greek, English
Github Skills (68)
coreference10
coreference-resolution10
nlp8
huggingface8
natural-language-understanding8
laravel7
python7
eloquent7
natural-language-processing7
mongodb7
machine-learning7
backend6
deep-learning6
react6
spacy6
Programming languages (6)
Github contributions (5)
dpasch01/spdx-compat-tools
Jun 2015 - Sep 2016
Contributions:10 commits, 9 pushes, 1 branch in 1 year 3 months
dpasch01/polarlib
Sep 2021 - Dec 2024
Code repository for POLAR, a framework for the modelling of polarization and identification of polarizing topics in news articles. POLAR utilizes a variety of computational method to model polarization, including NLP, Graphs, and Clustering. It produces hierarchical knowledge of polarization for a given dataset of news articles.
Contributions:2 PRs, 42 pushes, 3 branches in 3 years 3 months
Find and Hire Top DevelopersWe’ve analyzed the programming source code of over 60 million software developers on GitHub and scored them by 50,000 skills. Sign-up on Prog,AI to search for software developers.