Francesco Parisio
Staff Engineer at Deutsche Telekom
Berlin, Italy
Summary
👤
Senior🎓
Top SchoolFrancesco Parisio is a Staff Engineer based in Berlin who specializes in scaling distributed AI systems across thousands of NVIDIA GPUs, optimizing training, fine-tuning and inference for frontier LLMs and real-world products. With a PhD from EPFL and a decade of scientific computing experience, he blends academic rigor with hands-on production engineering to achieve large gains in throughput, latency and cost (e.g., 10x finetuning speedups and MFU >40% on 1000+ H100 clusters). He has built datacenter-scale inference stacks for >700B-parameter models, designed fault-tolerant architectures across Blackwell and Hopper generations, and led teams shipping satellite-AI and computer vision products. His background in applied physics and HPC informs deep expertise in NCCL profiling, DeepSpeed ZeRO-3, Ray/Kubernetes orchestration and GPU performance tuning. Known for turning complex research into operational systems, he also holds a track record of grant-funded research and international speaking engagements that reflect a non-linear career path bridging science and production.
9 years of coding experience
12 years of employment as a software developer
Ph.D., Computational Mechanics, Ph.D., Computational Mechanics at EPFLMaster Degree, Civil Engineering, Master Degree, Civil Engineering at Politecnico di Torino
Italian, French, English, German, Spanish
Github Skills (28)
deep-q-network10
openai-gym10
cross-entropy10
medium10
ml-agents10
reinforcement-learning9
hill-climbing9
python9
machine-learning9
statistics9
ddpg9
dqn9
markov-decision-processes9
dynamic-programming9
actor-critic9
Programming languages (3)
Github contributions (5)
fparisio/fparisio
Mar 2022 - Mar 2025
Contributions:13 pushes, 1 branch in 3 years
fparisio/mlops_churn
Sep 2022 - May 2023
Package for continuous deployment of customer churn
Contributions:12 PRs, 21 pushes, 8 branches in 7 months
continuous-deploymentdevopsdeploymentcontinuousdocker
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.