Benedikt S
Manager - Applied Research - NeMo Retriever at NVIDIA
New York, New York, United States
Summary
🤩
Rockstar🎓
Top SchoolBenedikt S is a manager and applied research leader with 9 years of experience building and scaling recommender systems and retrieval models, currently leading NeMo Retriever research at NVIDIA in New York. He combines hands-on deep learning engineering—particularly GPU-accelerated recommender pipelines—with team leadership, having progressed from individual contributor roles to managing applied research and deep learning teams. His open-source contributions to NVIDIA Merlin and NVTabular show practical expertise in productionizing end-to-end recommender workflows and adding robust statistical features for large-scale tabular feature engineering. A Columbia Data Science graduate with prior consulting and product ownership experience, he blends rigorous academic training with business-facing delivery across industries. Notably, he has driven measurable product impact (e.g., doubling CTR at Home24) and routinely bridges research, engineering, and deployment to move models into production.
9 years of coding experience
7 years of employment as a software developer
Master of Science Data Science, Master of Science Data Science at Columbia UniversityBachelor of Science (B.Sc.) Wirtschaftsmathematik, Bachelor of Science (B.Sc.) Wirtschaftsmathematik at University of Mannheim
German, English
Github Skills (13)
cudf10
machine-learning10
deeplearning-ai10
recommendation-system10
deep-learning10
tensorflow10
feature-engineering10
preprocessing10
gpu-acceleration10
python10
preprocess10
docker8
dockers8
Programming languages (7)
Github contributions (5)
NVIDIA-Merlin/Merlin
Mar 2021 - Dec 2022
NVIDIA Merlin is an open source library providing end-to-end GPU-accelerated recommender systems, from feature engineering and preprocessing to training deep learning models and running inference in production.
Role in this project:
ML Engineer
ML EngineerContributions:58 reviews, 21 commits, 18 PRs in 1 year 9 months
Contributions summary:Benedikt's contributions center around updating and modifying example notebooks related to recommender systems. They primarily worked on the "getting-started-movielens" examples, updating code and making edits related to data loading, preprocessing, and model inference using TensorFlow. Their changes involved integrating cuDF, updating docker files for model deployment, and modifying existing ETL processes with NVTabular, enhancing the examples with the necessary components for GPU acceleration and end-to-end recommender system pipelines.
in-productiontrainingdeep-learning-modelsengineeringtensorflow
NVIDIA-Merlin/NVTabular
May 2020 - Sep 2022
NVTabular is a feature engineering and preprocessing library for tabular data designed to quickly and easily manipulate terabyte scale datasets used to train deep learning based recommender systems.
Role in this project:Back-end Developer & Data Scientist
Back-end Developer & Data ScientistContributions:60 reviews, 75 commits, 51 PRs in 2 years 5 months
Contributions summary:Benedikt's commits focused on enhancing the `nvtabular` library's capabilities by adding statistical functions (mean, std, var) to the groupby operations, specifically for tabular data feature engineering. This involved modifying the `groupby.py` file to include these calculations, using `ddof=1` as the default setting for the degree of freedom. The user also added unit tests to verify these new functionalities.
tsneengineeringtensorflowpreprocessingnvidia
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.