Antreas Antoniou
Cofounder And Research Lead at Pieces
City of Edinburgh, Scotland, United Kingdom
Summary
🤩
Rockstar🎓
Top SchoolAntreas Antoniou is a machine learning researcher-engineer and cofounder with 11 years of experience building meta-learning and multi-modal systems, currently leading research at Axiotic AI and serving as Principal Research Partner at Pieces. He holds a PhD in Machine Learning from the University of Edinburgh and has combined academic rigor with industry impact through roles at Edinburgh, Google, and several AI startups. Antreas’s work spans few-shot/meta-learning (including maintaining the original "How to Train Your MAML" PyTorch replication) to on-device multi-agent ML for production products like Pieces OS and Copilot. He repeatedly focuses on practical data pipelines and efficient data loading—optimizing dataset handling and augmentation for few-shot benchmarks and courseware. Equally comfortable prototyping research and shipping engineering systems, he bridges research insights with production constraints to accelerate ML deployment. Based in Edinburgh, he pairs deep technical scholarship with entrepreneurial drive and a taste for reproducible, open-source tooling.
11 years of coding experience
9 years of employment as a software developer
Doctor of Philosophy - PhD Machine Learning, Doctor of Philosophy - PhD Machine Learning at The University of EdinburghMaster of Science (MSc) Data Science, Master of Science (MSc) Data Science at Lancaster University
Stackoverflow
58reputation
4kreached
3answers
2questions
Github Skills (18)
pytorch10
python10
machine-learning10
numpy10
data-preprocessing10
data-loading10
data-handling10
image-augmentation10
jupyter-notebook9
data-visualizations7
data-visualisation7
data-visualization7
keras-layer6
tensor6
backpropagation6
Programming languages (10)
Github contributions (5)
CSTR-Edinburgh/mlpractical
Sep 2017 - Jan 2020
Machine Learning Practical course repository
Role in this project:
ML Engineer
ML EngineerContributions:240 commits, 66 PRs, 210 pushes in 2 years 4 months
Contributions summary:Antreas's commits primarily involve modifications to the `mlp/data_providers.py` and `notebooks/01_Introduction.ipynb` files. They made changes to the data providers to improve efficiency and correct bugs within the code. Furthermore, the user updated the notebook introduction file, addressing issues and including a graph visualization of the data. These changes reflect a focus on improving data handling and visualization for a machine learning course.
machine-learningneural-networksjupyter-notebook
AntreasAntoniou/HowToTrainYourMAMLPytorch
Oct 2018 - Nov 2021
The original code for the paper "How to train your MAML" along with a replication of the original "Model Agnostic Meta Learning" (MAML) paper in Pytorch.
Role in this project:ML Engineer
ML EngineerContributions:1 review, 97 commits, 6 PRs in 3 years 1 month
Contributions summary:Antreas primarily contributed to the data loading and preprocessing pipeline for a few-shot learning system. They modified the `data.py` file to implement image augmentation and rotation, including the integration of a new class for image rotation. The user also worked on improving the dataset loading process, including changes to handle mini-imagenet datasets and adjustments to the seed handling for unique tasks. Moreover, they optimized data loading by integrating the loading of image batches and incorporating a mechanism for dynamic dataset loading into RAM.
mamlmeta-learningpytorchreplication
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.