Dr. Artiom Kovnatsky

• Specializing in “data activation”, I provide hands-on expertise and consultancy across industries, using creativity, machine learning, business analysis, and MLOps to unlock value from (complex and “imperfect”) data. • Managed and delivered high-risk projects, ensuring on-time completion and high-quality outcomes. • Successfully applied for grants, proposing innovative AI research projects. See more details: www.artiomkovnatsky.com

• Developed cutting-edge data science products for clients, focusing on vehicle data analytics, including projects for Skoda GmbH.

• Established and led the data science department, building a decentralized team from the ground up. • Defined the vision, strategy, and tech stack, ensuring timely delivery of data-driven solutions in the real estate domain.

• Developed survival models for predictive maintenance, optimizing machine maintenance schedules in collaboration with a specialized team.

• Consulted the CEO on developing a predictive analytics engine for HR data. • Authored a successful grant proposal, securing €1.2M in funding for the project. • Provided critical last-minute project support, delivering solutions under tight deadlines.

• Developed the PREISATLAS (https://www.homeday.de/de/preisatlas) real estate price map, resulting in a 1500% increase in traffic and a 750% increase in qualified leads. • Created the HOMEDAY-estimation package, a core tool for automatic property evaluation used daily by the company. • Proposed and presented innovative business models, influencing company strategy. • Enhanced business processes through data-driven insights. • Presented work at conferences, consistently recognized as one of the top speakers.

• Developed a prototype evaluation engine for cadastral land units, forming the foundation for Estonia’s 2020-2021 land tax reform.

• Consulted on data science and business strategy for a new product launch.

• Conducted advanced research under Prof. Michael M. Bronstein, with work published in top-tier journals and conferences. • Delivered presentations across various international venues. • Taught Computer Vision and Mathematics courses. Google scholar: https://tinyurl.com/3rz8mjyr

• Research & development in Image Processing: Technicolor submitted a patent.

Thesis: Spectral Methods for Multimodal Data Analysis Supervisor: Prof. Michael M. Bronstein Google scholar: https://tinyurl.com/3rz8mjyr

Manifold Learning