Pavel Esir
Software Engineer at Intel Corporation
Contact Info
Other
Summary
🤩
Rockstar🎓
Top SchoolPavel Esir is a software engineer with 12 years of experience specializing in back-end development and machine learning engineering, currently working at Intel. He has contributed substantially to OpenVINO, improving the model optimizer with mixed-precision support, integer-type handling, new operation support and robust test coverage for a widely used AI inference toolkit. His work spans model conversion and optimization for practical deployment, including adapting large models (DeepFloyd T5, RedPajama) and integrating voice and grounded segmentation models in notebook tutorials. Pavel’s background in academic research and teaching informs a disciplined, test-driven approach to engineering complex ML tooling. He combines production-grade system engineering with deep familiarity in model optimization pipelines, making him effective at bridging research prototypes and industrial inference deployments.
12 years of coding experience
3 years of employment as a software developer
N.I. Lobachevsky State University of Nizhni Novgorod
Github Skills (19)
back-end-development10
python10
machine-learning10
inference10
openvino10
deep-learning10
optimisation10
model-optimization10
computer-vision10
model-conversion10
optimization10
c-language9
ml9
cprogramming-language9
quantization9
Programming languages (4)
Github contributions (5)
openvinotoolkit/openvino
May 2020 - Jan 2023
OpenVINO™ is an open source toolkit for optimizing and deploying AI inference
Role in this project:
Back-end Developer & ML Engineer
Back-end Developer & ML EngineerContributions:1184 reviews, 86 commits, 191 PRs in 2 years 8 months
Contributions summary:Pavel's contributions centered around improving and extending the model optimizer, a critical component for OpenVINO, including implementing a more precise handling of integer types, supporting mixed-precision inference for quantized IRs by enhancing and extending the core logic of the framework, and adding support for various operations (e.g., Gather, StridedSlice, Convolution). The user implemented fixes, refactoring, and optimizations. Furthermore, the user worked on new API's, creating unit-tests, and adding tests for various functionalities to the model optimizer.
inference-enginepytorchmodel-optimizerdeep-learninggpu
openvinotoolkit/openvino_notebooks
Oct 2023 - Mar 2024
📚 Jupyter notebook tutorials for OpenVINO™
Role in this project:ML Engineer
ML EngineerContributions:18 reviews, 7 PRs, 62 comments in 5 months
Contributions summary:Pavel's commits primarily focus on integrating and optimizing models for inference within the OpenVINO framework, demonstrating a strong emphasis on deep learning and computer vision. They worked on adapting and refining model upcasting techniques (e.g., FP32 calibration) to improve performance and accuracy for models like DeepFloyd T5, RedPajama, and Tiny SD. The contributions include converting models to OpenVINO IR format and resolving compatibility issues to ensure proper functioning within the notebook environment. They also worked on integrating the OpenVoice model for voice cloning and GroundedSAM models.
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.