Nikita Savelyev
Munich, Bavaria, Germany
Summary
🤩
RockstarNikita Gusev is an AI Frameworks Engineer based in Munich with nine years of experience advancing high-performance computing and AI/ML/DL stack performance at Intel. He specializes in low-level C/C++ development for MPI and oneCCL, optimizing collective communications and memory efficiency across CPU/GPU fabrics while working with SYCL, TensorFlow, PyTorch and Horovod. In open source he has contributed to OpenVINO’s NNCF and notebook ecosystem, adding accuracy-aware quantization and INT8 demos for models like Whisper and Stable Diffusion that translate systems work into measurable inference gains. His background—growing from MPI/OFI intern to framework engineer—reflects a rare blend of network-level expertise and hands-on model optimization that accelerates real-world AI workloads.
9 years of coding experience
Github Skills (14)
transformers10
quantization10
compression10
tensorflow210
machine-learning10
pytorch10
inference10
deep-learning10
tensorflow10
python10
openvino10
nlp9
pruning9
computer-vision9
Programming languages (3)
Github contributions (5)
openvinotoolkit/nncf
Nov 2022 - Jan 2023
Neural Network Compression Framework for enhanced OpenVINO™ inference
Role in this project:
ML Engineer
ML EngineerContributions:416 reviews, 10 commits, 127 PRs in 2 months
Contributions summary:Nikita made several contributions related to the Neural Network Compression Framework (NNCF). Their work involved refactoring and improving the compression process, including making modifications to handle compression stage and rate implementations. They focused on optimizing the adaptive compression training loop and fixed bugs in the existing code. They also enhanced the accuracy-aware training loop by changing function signatures.
bertsemantic-segmentationmixed-precision-trainingtensorflowclassification
openvinotoolkit/openvino_notebooks
Mar 2022 - Mar 2022
📚 Jupyter notebook tutorials for OpenVINO™
Role in this project:ML Engineer
ML EngineerContributions:91 reviews, 1 commit, 31 PRs in 1 day
Contributions summary:Nikita contributed to the development and optimization of machine learning models within the OpenVINO notebooks repository. Their work focused on integrating and comparing NNCF quantization techniques within existing notebooks, specifically for models like Whisper, Dolly 2.0, Stable Diffusion v2, and other text-to-image generation pipelines. They made modifications to code, added and refined calibration datasets, and implemented performance comparisons between FP32 and INT8 models. They also incorporated interactive demos to show the benefits of the model optimization.
deep-learningjupyter-notebooknotebookinferencecomputer-vision
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.