Jan Ebert
CFO at Volkswagen Financial Services | U.S.
United States
Summary
🤩
Rockstar🎓
Top SchoolJan Ebert is a finance executive and CFO with over a decade of leadership at Volkswagen Financial Services, steering controlling, strategy, M&A and investment management across international markets. He combines corporate finance rigor—rooted in a CFA and a master's in finance—with hands-on transaction and supervisory board experience across mobility and tech-adjacent businesses. Notably, Jan pairs this financial leadership with practical technical involvement in open-source ML and robotics projects, contributing DeepSpeed integrations to a popular DALL-E PyTorch replication and fixes/features to PyTorch and Flux.jl. That technical thread reflects a rare mix of quantitative, operational, and engineering fluency, enabling data-driven decision making for complex M&A and strategic programs. Currently based in the United States, he is completing executive education at HEC Paris to further scale enterprise and cross-border finance capabilities.
10 years of coding experience
5 years of employment as a software developer
Chartered Financial Analyst, Finance, General, Chartered Financial Analyst, Finance, General at CFA InstituteMaster's degree, Finance, General, Master's degree, Finance, General at University of HanoverExecutive Education, Executive Education at HEC Paris
Github Skills (34)
transformers10
unit-testing10
pytorch10
artificial-intelligence10
distributed-training10
c-language10
robotics10
python10
artificial-neural-networks10
machine-learning10
deeplearning-ai10
moveit10
deepspeed10
deep-learning10
flux10
Programming languages (15)
Github contributions (5)
lucidrains/DALLE-pytorch
Mar 2021 - Sep 2021
Implementation / replication of DALL-E, OpenAI's Text to Image Transformer, in Pytorch
Role in this project:
MLOps Engineer
MLOps EngineerContributions:19 reviews, 55 commits, 28 PRs in 6 months
Contributions summary:Jan primarily focused on integrating DeepSpeed, a distributed training library, into the DALL-E PyTorch implementation. Their contributions included adding DeepSpeed utility functions, integrating DeepSpeed support into training scripts (VAE and DALL-E), and ensuring correct checkpointing and model loading with DeepSpeed. Additionally, the user implemented features related to distributed data loading and half-precision training, further enhancing the training process.
pytorchmulti-modaltransformersdall-eattention-mechanism
FluxML/Flux.jl
Aug 2019 - Apr 2020
Relax! Flux is the ML library that doesn't make you tensor
Role in this project:ML Engineer
ML EngineerContributions:25 commits, 5 PRs, 58 comments in 7 months
Contributions summary:Jan primarily contributed to the Flux.jl machine learning library by fixing bugs, improving documentation, and adding tests. Their work involved correcting CuArrays imports and fixing binary crossentropy on CuArrays, indicating a focus on GPU support. Furthermore, the user implemented and improved the library's utility and loss functions, as well as added and updated the documentation.
ml-librarythe-human-braindata-sciencedeep-learningneural-networks
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.