Huu4Ontocord is the CEO of Ontocord.AI with 11 years of hands-on experience in machine learning engineering and data tooling. He blends executive leadership with deep technical contributions, notably to high-profile open-source projects like Hugging Face Datasets and the BigScience workshop. His work on ConceptNet5 and modifications to NELL, LAMA, Ollie and Atomic datasets shows a specialization in common-sense reasoning and knowledge-base curation. He has also improved data access layers and experimented with model attention mechanisms (AliBi) while adapting GPT-2, indicating a strong end-to-end ML systems orientation. Comfortable moving between strategic product decisions and low-level code, he uniquely combines CEO responsibilities with active technical authorship. Colleagues can expect a leader who still dives into dataset engineering and model plumbing to drive product-forward research.
🤗 The largest hub of ready-to-use datasets for ML models with fast, easy-to-use and efficient data manipulation tools
Role in this project:
Data Scientist
Contributions:24 reviews, 5 commits, 7 PRs in 26 days
Contributions summary:Huu4Ontocord contributed to the `huggingface/datasets` repository by adding and modifying datasets related to common sense reasoning and knowledge bases. Their work involved creating a new dataset for ConceptNet5 and making changes to existing datasets such as NELL, LAMA, Ollie and Atomic by modifying the code, documentation, and related files. The user demonstrated their involvement in managing datasets, along with making fixes and other updates within the datasets.
Central place for the engineering/scaling WG: documentation, SLURM scripts and logs, compute environment and data.
Role in this project:
ML Engineer
Contributions:14 commits, 14 pushes, 1 comment in 3 months
Contributions summary:Huu4Ontocord primarily focused on extending and modifying the `datastore.py` file, which appears to provide optimized data storage and access functionalities for datasets, including support for SQL databases, memmap files, and indexed gzip compression. They introduced and refined classes related to data access and manipulation, indicating efforts to improve how data is handled within the repository. Furthermore, the user added a copy of `modeling_gpt2.py` to experiment with the addition of AliBi attention, showing focus on adapting a GPT-2 model for a specific project need.
nlpplaceslurmmodelsmachine-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.