Projects

Snippet — AI Video Learning Platform

02 — Independent

Snippet — AI Video Learning Platform

Snippet transforms educational video into a personalized, searchable learning experience. It transcribes and intelligently tags creator content so a learner can search a concept like "how to build microservices" and instantly get timestamped excerpts from across many videos, complete with previews and the ability to assemble a custom learning plan.

Key features

  • AI transcription and tagging pipeline. On upload, Snippet transcribes video (Whisper or AWS Transcribe), extracts key topics, tools, and concepts with entity recognition and semantic analysis, and stores weighted, timestamp-linked transcript segments for deep indexing.
  • Semantic video search. Natural-language queries (e.g., "how to deploy with Docker") return timestamped snippets from across the library, ranked by relevance.
  • Snippet previews and learning plans. Each result includes a short preview aligned to the query; learners can open the full video or save the snippet to a personalized learning path.
  • Modular, trackable journeys. Saved snippets organize into themed journeys with progress tracking and recommendations for what to learn next.

How it works

  1. Ingest a video and transcribe it to a timestamped transcript.
  2. Run entity and concept extraction to tag segments with tools and topics (Docker, AWS, Golang, and more), weighted by relevance.
  3. Index timestamp-linked segments for semantic retrieval.
  4. At query time, embed the query, retrieve the most relevant segments across videos, and return ranked, previewable snippets.

Why it matters

Snippet makes high-quality learning modular and measurable — closer to building a playlist than watching a course — by indexing knowledge at the concept level instead of the video level.