TIME SERIES · DEEP LEARNING
InceptionTime
An ensemble of deep convolutional networks for scalable time series classification, with an open-source implementation of the research.
Explore the code
Hassan Ismail FawazRESEARCH & ENGINEERING
Building on research to solve problems with machine learning.
PRODUCT · CREATOR
AI coaching for Rocket League.
I created GetBetter to turn Rocket League replay data into actionable coaching. It analyzes key moments in a match, offers feedback on kickoffs, shooting, saves, and recovery, and helps players track their progress.
Visit GetBetterOPEN RESEARCH
TIME SERIES · DEEP LEARNING
An ensemble of deep convolutional networks for scalable time series classification, with an open-source implementation of the research.
Explore the codeRESEARCH · REPRODUCIBILITY
An open-source benchmark accompanying our review: 8,730 trained models evaluated across 97 time series datasets.
Explore the benchmarkTRANSFER LEARNING · BENCHMARKS
A benchmark of deep unsupervised domain adaptation methods for time series classification.
Explore the benchmarkMORE OPEN-SOURCE TOOLS
OPEN-SOURCE CONTRIBUTIONS
I authored the official Keras example for training a convolutional time series classifier on the FordA dataset, from preparing the data to evaluating the model.
Read the Keras tutorialI contributed integrations for three text classification datasets to the Hugging Face datasets library. Each contribution was merged upstream.
INDUSTRY
Bringing advances in generative AI to the General Organization for Social Insurance.
Fine-tuning large language models, benchmarking vector databases, and improving retrieval-augmented generation.
Applying artificial intelligence to wireless telecommunication networks.
Developing machine learning solutions for automatic content moderation.