Paul Almasan
Researcher at Telefónica Innovación Digital
Barcelona, Catalonia, Spain
Summary
👤
Senior🎓
Top SchoolPaul Almasan is an applied AI researcher with 11 years of experience and 6+ years focused on end-to-end machine learning for wireless networks, currently developing graph neural network and digital-twin solutions at Telefónica Innovación Digital in Barcelona. He designs scalable models for radio coverage prediction as practical alternatives to ray tracing and applies generative AI to large-scale telecom datasets. His PhD work produced influential GNN-based routing and DRL architectures that attracted academic citations and open-source interest, and he has led coordination of multi-partner EU projects (INSTINCT, 6G-MIRAI). Comfortable bridging academia and industry, he mentors MSc/PhD students and has hands-on experience implementing demos, reproducible experiments, and production-ready research prototypes. An under-the-radar strength is his mix of time-series compression and spatio-temporal analysis expertise from collaborative visiting research stays, which complements his network optimization focus.
11 years of coding experience
3 years of employment as a software developer
UPC Universitat Politècnica de CatalunyaErasmus exchange in the Bachelor of Information Technologies program, Information Technology, Erasmus exchange in the Bachelor of Information Technologies program, Information Technology at České vysoké učení technické v Praze
Spanish, Catalan, Romanian, English, Japanese, French
Github Skills (16)
graph-neural-network9
5g9
raytracing9
reproducible-research8
communications7
machine-learning7
algorithm6
deep-learning6
computer-networks6
gpu-acceleration6
reinforcement-learning6
open-source5
deep-reinforcement-learning4
machine-learning-models3
wan3
Programming languages (2)
Github contributions (5)
knowledgedefinednetworking/DRL-GNN
Nov 2021 - Jan 2023
Contributions:29 commits, 6 PRs, 27 pushes in 1 year 2 months
BNN-UPC/ENERO
Jul 2022 - Jan 2023
Code used in the paper "ENERO: Efficient real-time WAN routing optimization with Deep Reinforcement Learning". In this paper, the DRL agent is implemented with the PPO algorithm
Contributions:15 commits, 1 PR, 14 pushes in 6 months
deep-reinforcement-learningwandeep-learninggraph-neural-networksmachine-learning
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.