Sherif Akoush
Cambridge, England, United Kingdom
Summary
🤩
Rockstar🎓
Top SchoolSherif Akoush is a seasoned MLOps and ML engineer with a PhD in Computer Science from the University of Cambridge and over a decade building production ML systems. He has led LLM/ML serving platforms at Seldon and now heads Simulation Intelligence Ops, bringing deep experience in model serving, inference optimization, and production pipelines. Sherif is an active open-source contributor to prominent projects like Seldon Core and MLServer, where he improved environment handling, MLflow v2 support, DataFrame output encoding and added explainability runtimes—work that directly improves scalability and usability for thousands of deployed models. His background spans academia and industry, from country-scale telecom analytics to cloud-native ML products for finance and automotive, which gives him a rare combination of research rigor and production focus. Comfortable across the stack, he focuses on maintainable utilities, robust testing, and pragmatic integrations that reduce friction for ML teams. Based in Cambridge, he blends systems-thinking with hands-on engineering to move complex ML workflows from prototype to reliable, auditable production.
10 years of coding experience
12 years of employment as a software developer
Doctor of Philosophy (PhD) Computer Science, Doctor of Philosophy (PhD) Computer Science at University of Cambridge
Github Skills (21)
kubernetes10
deploying10
python10
pandas10
machine-learning10
mlflow10
numpy10
mlops10
kubernetes-pods10
testing9
pytorch9
pytest9
docker8
scikit8
scikit-learn8
Programming languages (9)
Github contributions (5)
SeldonIO/MLServer
Aug 2021 - Jan 2023
An inference server for your machine learning models, including support for multiple frameworks, multi-model serving and more
Role in this project:
ML Engineer
ML EngineerContributions:249 reviews, 13 commits, 227 PRs in 1 year 5 months
Contributions summary:Sherif primarily contributed to the `mlserver-mlflow` runtime, enabling support for pandas DataFrames and Series in the output encoding for MLflow models. They modified the `to_outputs` function to handle DataFrame and Series conversion to NumPy arrays for compatibility. Further contributions included adding tests for the new encoding capabilities, integrating PyTorch into the dev dependencies, and fixing linting issues. They also pinned the version of fastapi. The user also added a new alibi-explain runtime, including adding kernel shap, getting names from metadata endpoints, and making sure the output is correct.
xgboostmulti-frameworkspythonseldon-corelightgbm
SeldonIO/seldon-core
Jul 2021 - Jan 2023
An MLOps framework to package, deploy, monitor and manage thousands of production machine learning models
Role in this project:MLOps Engineer
MLOps EngineerContributions:604 reviews, 99 commits, 722 PRs in 1 year 6 months
Contributions summary:Sherif primarily focused on refactoring and enhancing the environment variable retrieval process within the `seldon-core` repository, specifically related to model and image names. They moved environment variable handling to a dedicated utility module and added unit tests to ensure the robustness of these changes. Furthermore, the user contributed to supporting the MLFlow v2 protocol, including integration tests and end-to-end examples, demonstrating their involvement in model serving and deployment. This suggests a focus on improving the usability, maintainability, and functionality of the MLOps framework.
monitorservingdata-sciencedeploymentmachine-learning
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.