Paolo Toccaceli
Technical Specialist at Advanced Research + Invention Agency (ARIA)
United Kingdom
Summary
👤
Senior🎓
Top SchoolPaolo Toccaceli is a Technical Specialist with 11 years of experience applying machine learning and systems engineering across AI research, VLSI-style compute architectures, and telecom systems. He has transitioned from embedded and firmware engineering to leading ML model optimization at Graphcore, where he scaled CV and NLP models on novel hardware using TensorFlow, PyTorch and occasional C++ fixes. His work bridges research and product: academic projects on calibrated probabilistic prediction and conformal change-point detection (Amazon-funded) inform practical implementations for drug discovery and large-scale model training. At ARIA he now focuses on technical innovation at the intersection of advanced computing and applied research. Paolo’s background in radio systems and capacity engineering gives him an uncommon systems-level perspective on performance, parallelism and real-world deployment constraints. He holds advanced degrees in data science and is pursuing a part-time PhD in machine learning, combining deep academic rigour with production-focused engineering.
11 years of coding experience
30 years of employment as a software developer
Doctor of Philosophy - PhD (part-time) Machine Learning, Doctor of Philosophy - PhD (part-time) Machine Learning at Royal Holloway, University of LondonMSc in Electronic Engineering (Computer Science), MSc in Electronic Engineering (Computer Science) at Politecnico di Milano
English, Italian
Github Skills (50)
data-analysis10
python10
data-science10
spyder10
statistics10
ide10
machine-learning10
anaconda10
scikit-learn10
qt9
python-development9
language-server-protocol9
react-component8
jupyter8
notebook7
Programming languages (6)
Github contributions (5)
ptocca/VennABERS
May 2020 - Nov 2022
Fast implementation of Venn-ABERS probabilistic predictors
Contributions:7 commits, 16 pushes, 3 branches in 2 years 6 months
ptocca/ILSVRC2012_CP
May 2020 - Aug 2024
Applying Inductive Conformal Predictors to ILSVRC2012 with ResNet50
Contributions:64 pushes, 2 branches in 4 years 3 months
deep-learningpytorchresnet50predictors
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.