Ekaterina “Kate” Deyneka is founder and CEO of Reelful, an agentic video editor. A former Snapchat machine-learning engineer and applied researcher, she applies code-generation tools to editing real photos and footage.[1][2]
Start with Building an Agentic Video Editor for Mass Consumer →
GitHub 
X Career
Deyneka studied applied mathematics and physics at Moscow Institute of Physics and Technology, pursued graduate work in nanobiology at Delft and Erasmus Medical Center, and conducted computer-science research at UC Irvine. At Snapchat she worked on image and video generation. Her independent work includes a PDF-reading extension and speech-feedback tools. Reelful applies her research and interface work to short-form video creation.[1]
Contributions
Reelful first analyzes uploaded media and speech, then asks the user to approve a creative plan. An agent operates in a remote sandbox and writes a React/Remotion composition, arranging real footage with supporting captions, animation, music or narration. Verification checks whether the composition renders and returns failures for revision. Chat and manual controls let the user change direction. The approach adapts a coding-agent loop to an artifact whose output is a video.[3]
Tools, writing & talks
Writing, research & talks
Start here
A technical account of media analysis, approved plans, code-based compositions and render checks.
Watch the architecture and verification chapters to see how an edit becomes a code artifact. Basic knowledge of coding agents helps; the product demonstration is accessible without React experience.
Speaker[3]An agentic video editor built like an agentic app builder: raw footage and a line of direction go into a sandbox, skills carry the cut rules and typography, a code-based composition layer renders the result, and a verification pass catches broken output.
Catalog workOriginal source