Mikhail Kozhan is engaged in developing solutions in the fields of artificial intelligence, machine learning, and business process automation. With over 15 years of experience working with IT systems, data, and complex business challenges, he has successfully implemented projects ranging from requirements analysis to the launch of finished systems. His specialization includes the creation of AI services, RAG systems, ML models, B2B platforms, and internal web solutions. He develops his own products, such as ML Platform and Feature Hub, and actively utilizes technologies like Python, Docker, Kubernetes, and LLM. He operates on a full-cycle principle, overseeing everything from architecture to implementation and operation.

Reviews

Mikhail Kozhan is a professional and responsible freelancer who delivers high-quality work on time and with a high level of engagement. Reviews highlight his responsiveness, willingness to make additional adjustments, ability to propose solutions, and provide consultations on design and technical matters. There are no negative comments in the reviews.

Portfolio projects

Skills and Technologies: web development, system design, content management, machine learning, PostgreSQL, API, integration with external systems.

Projects and Achievements: Mikhail has designed and developed three key platforms: Content Factory for content management, Feature Store for centralized feature preparation, and ML Platform for managing the lifecycle of ML models. In the Content Factory project, a system was created that integrates topic search, editorial processing, and publication into a single interface, enhancing process visibility and simplifying content management. The Feature Store ensured reproducibility and data quality control, helping to avoid duplication and improve collaboration among specialists. The ML Platform combined model versioning and artifact management, providing state control and integration with corporate infrastructure.

Achievements and Recognition: All three platforms were developed as standalone software products with a focus on reproducibility, resilience, and integration. The outcomes include the establishment of managed processes, significantly increasing efficiency in handling content and data, as well as ensuring secure access and version control, which is crucial for corporate use.

* It is created by a neural network based on a portfolio, information provided by the user about himself and reviews from other users.

русская версия