Костюков Максим is a software engineer and FEFU student based in California with seven years of hands-on experience in ML engineering and DevOps. He contributes to the popular OPHoperHPO image-background-remove-tool, where he refactored architecture by integrating FastAPI, hardened Docker/CUDA workflows, and added fp16 tests to boost model performance. His work shows a practical focus on production readiness—improving build efficiency, model caching, and API security (SSR F filter fixes and User-Agent handling). Comfortable bridging research models and deployment, he optimizes both inference performance and operational reliability. Despite being a student, he operates at the intersection of ML and infrastructure, delivering measurable improvements to open-source ML tooling.
✂️ Automated high-quality background removal framework for an image using neural networks. ✂️
Role in this project:
DevOps & ML Engineer
Contributions:9 releases, 5 reviews, 45 commits in 8 months
Contributions summary:Костюков refactored the project's architecture by integrating FastAPI and splitting the requirements. They focused on improving the build process by fixing the Dockerfile for efficiency and addressing CUDA/non-CUDA environment separation. Furthermore, the user added tests for the fp16 configuration, demonstrating work related to optimizing the model for performance, and also implemented the integration of a new caching system for the models. They also worked on fixing the API ssrf filter malfunction and added User-Agent header to the model downloader.
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.