Price: 0
Number of applications: 1
16.12.25 (inclusive)
By agreement
Finished product
ICT tasks
Media sphere
Neurotechnology and artificial Intelligence
Software/ IS,
Mobile app
Students often do not review long video lectures, even when they need to repeat the material. Problems: • lessons are 10-25 minutes long • it's difficult to find the right moment quickly • Low “repetition” efficiency • students prefer short discussions in TikTok format • it is impossible to manually create mini-videos for all lessons — it is too long and expensive We need an automatic tool that cuts out the key points by itself and makes short, convenient mini-explanations.
✔ automatic short video generator that does not require manual installation ✔ New content formats that increase retention ✔ Mini-showdowns that are great for teenagers and their habits ✔ Save teachers' time ✔ the ability to create TikTok/Shorts/Reels content automatically ✔ improved repetition of materials (spaced repetition)
Nurlan Rakhimzhanov
Purpose and description of task (project)
Create a module that automatically converts long video tutorials into short, understandable and structured mini-videos (30-90 seconds), highlighting key points, formulas, definitions and examples. This will allow students to repeat topics faster and improve engagement. Within the framework of the project, it is necessary to implement: ⸻ A. Extracting key points from a video Usage: • Speech recognition (Whisper/Vosk/Speech-to-Text API) • NLP/LLM to identify important fragments • determining peaks of importance by words: “formula", “main", “example", “rule” ⸻ B. Mini video generation • automatic cutting of necessary fragments • combining several pieces into a 30-90 second structure • automatic cropping and rippling of the format 9:16 / 16:9 / 1:1 • automatic insertion of titles and subtitles • Smooth transitions and visual effects ⸻ C. Inserting AI-generated hints • Explanations • key formulas • Important definitions • a short cheat sheet on the screen ⸻ D. Server side (Backend + Worker) • Video processing via FFmpeg • NLP module for defining semantic segments • Mini-timeline generator • queue system (RabbitMQ/Kafka) • storage of mini-videos on CDN/S3 ⸻ E. Embedding in a mobile application • Showing mini-videos inside lessons • separate “Quick topic analysis” feed • auto-generation of a selection of mini-videos • Caching and offline access ⸻ F. Documentation • API description • processing scheme • recommendations for teachers on how to improve the quality of mini-videos
Whatsapp
