10 Hugging Face machine learning engineers
Library and tooling work shows up directly in the roles of these Hugging Face machine learning engineers: Régis Pierrard maintains Optimum, Harry Mellor maintains vLLM, Simon Brandeis built AutoTrain, Lewis Tunstall builds RLHF tooling and Nouamane Tazi trains LLMs. Nine are still at Hugging Face and six are based in France, so they suit a team deciding how to optimize, train or serve transformer models.
Professionals to explore
Anton Lozhkov
LinkedInExperience: Machine Learning Engineer · Hugging Face
$336 per 30-minute call
Anton Lozhkov describes a focus on making recent NLP research more efficient and usable in real-life scenarios, with interests in few/zero-shot learning, transformer models for NLP and CV, and ML infrastructure, as a Machine Learning Engineer at Hugging Face since 2021. Anton was Senior Machine Learning Engineer at Embedika (2020-2021) and a Senior Data Engineer at NAUMEN (2019-2020).
Is this you? Claim your profile and get paid for calls.
Carl Cochet
LinkedInExperience: Former Machine Learning Engineer · Hugging Face
$336 per 30-minute call
Carl Cochet was a Machine Learning Engineer at Hugging Face from 2022 to 2023 and has been a Data Engineer at Invyo since 2023. Carl was a Data Scientist at Euranova (2020-2022) and did an end-of-studies data science project at Airbus (2018-2019), the one here who can compare Hugging Face engineering with data engineering work elsewhere.
Is this you? Claim your profile and get paid for calls.
Harry Mellor
LinkedInExperience: Machine Learning Engineer · Hugging Face
$370 per 30-minute call
Harry Mellor is a Machine Learning Engineer at Hugging Face (since 2025) and a Maintainer of vLLM (since 2025), after Contributor to vLLM (2024-2025). Harry was Senior AI Engineer (2024) and AI Engineer (2022-2023) at Graphcore and holds an MEng in Engineering Science from the University of Oxford, so serving models with vLLM is the natural topic.
Is this you? Claim your profile and get paid for calls.
Lewis Tunstall
LinkedInExperience: Machine Learning Engineer · Hugging Face
$1,346 per 30-minute call
On the Hugging Face research team, Lewis Tunstall builds tools and educational content for reinforcement learning from human feedback (RLHF), as a Machine Learning Engineer since 2021. Lewis was Senior Data Scientist at Swisscom (2020) and a Lecturer at Berner Fachhochschule BFH (2019-2020), a fit for questions on RLHF tooling and on teaching teams to use it.
Is this you? Claim your profile and get paid for calls.
Morgan Funtowicz
LinkedInExperience: Machine Learning Engineer · Hugging Face
$336 per 30-minute call
Morgan Funtowicz has been a Machine Learning Engineer at Hugging Face since 2019, after Research Engineer at NAVER LABS Corp. (2018-2019), and lists open-source contributions to Microsoft's CNTK and Facebook's PyTorch. Morgan cites machine reading and comprehension, and control and planning methods in distributed environments, and was a Research Intern at Microsoft Research (2016-2017), a fit for questions on machine learning in distributed environments.
Is this you? Claim your profile and get paid for calls.
Nouamane Tazi
LinkedInExperience: Machine Learning Engineer · Hugging Face
$471 per 30-minute call
Nouamane Tazi describes the Machine Learning Engineer role at Hugging Face (since 2022) as training LLMs and "making GPUs go brrr", and was ML Research Engineer Intern there (2022). Nouamane was also Software Engineer & Data Scientist Intern at Flashbrand Inc. (2020-2021); the natural topic is GPU efficiency when training large language models.
Is this you? Claim your profile and get paid for calls.
Pablo Montalvo
LinkedInExperience: Machine Learning Engineer · Hugging Face
$336 per 30-minute call
Pablo Montalvo, who holds a PhD in physics, has been a Machine Learning Engineer at Hugging Face since 2023, after Lead Data Scientist (2021-2023) and Research Scientist (2018-2023) at Rakuten France. Pablo was also Visiting Research Scientist at Rakuten (2017-2018) and Postdoctoral Research Scientist at Tainan Hydraulics Laboratory (2015-2016), useful for questions on moving from ML research at Rakuten to engineering at Hugging Face.
Is this you? Claim your profile and get paid for calls.
Régis Pierrard
LinkedInExperience: Machine Learning Engineer · Hugging Face
$336 per 30-minute call
Régis Pierrard has been a Machine Learning Engineer on Hugging Face's optimization team since 2022, as a maintainer of Optimum, an open-source library. Régis was AI Engineer at AnotherBrain (2020-2022) and a PhD student in Explainable Artificial Intelligence at CEA (2017-2020), a fit for questions on model optimization with Optimum and on explainable AI.
Is this you? Claim your profile and get paid for calls.
Simon Brandeis
LinkedInExperience: Machine Learning Engineer · Hugging Face
$303 per 30-minute call
Simon Brandeis built, deploys and maintains AutoTrain as a Machine Learning Engineer at Hugging Face (since 2020), and describes experience building ML applications and their cloud infrastructure. Simon was Data Scientist at Sancare (2020) and Python Software Engineer at Arch Systems (2019), useful for questions on AutoTrain and on deploying ML applications in the cloud.
Is this you? Claim your profile and get paid for calls.
Tom Aarsen
LinkedInExperience: Machine Learning Engineer · Hugging Face
$673 per 30-minute call
Tom Aarsen holds Machine Learning Engineer (since 2023) and Hugging Face Fellow (since 2023) titles at Hugging Face, and was a Machine Learning Engineer at Argilla (2023) and Data Scientist at Squadra Machine Learning Company (2022). The Argilla and Hugging Face engineering roles make Tom a fit for questions on open-source machine learning engineering work.
Is this you? Claim your profile and get paid for calls.
Choose the right perspective
For model optimization, Régis Pierrard (Optimum) is the match, and Harry Mellor for vLLM; Nouamane Tazi covers LLM training on GPUs and Lewis Tunstall RLHF tooling. Simon Brandeis suits AutoTrain questions, Anton Lozhkov and Morgan Funtowicz research-to-production NLP, and Carl Cochet, who left in 2023, a view from outside the company.
Questions to take into the conversation
- 01How should a team decide whether to optimize a transformer model with a library like Optimum or serve it as-is with vLLM?
- 02What tooling and data does an RLHF run need beyond a standard training setup?
- 03Where do GPU bottlenecks usually show up when training an LLM with open-source libraries?
About this directory
This is a professional research starting point based on business profile data retrieved on . Titles and companies reflect that source snapshot; roles marked former had ended by then. Check the linked profiles for current details. Inclusion does not imply Instant Expert membership or availability.