Trupeer: Video-First Knowledge Systems
Founding engineer at Trupeer. Built core video, documentation, and search systems: transcription pipelines, low-latency retrieval, AI features for docs, translation, voiceovers, and avatars.
The Challenge
Take raw screen recordings, the way people actually capture knowledge, and turn them into searchable, navigable, multilingual content. Videos, documentation, knowledge bases, avatar-driven workflows. All of it built for real B2B usage.
What I Built
- Pipelines that turned raw recordings into videos, structured documentation, and searchable knowledge, including transcription, indexing, and retrieval (docs.trupeer.ai is one example surface).
- Search across video and document content, letting users find specific moments in long recordings via transcript-anchored navigation. Custom low-latency transcription and indexing pipeline made this work in real time.
- AI features: document generation from videos, translation, voiceovers, avatar-based video workflows. Stitched together with the rest of the platform so that one upload could produce a fan of artifacts in different languages, formats, and presenters.
- Knowledge base and content organization systems, plus performance and reliability work for real production usage across teams.
A specific moment
Built a video-first knowledge system for B2B clients in under two days that generated $150K in revenue within a week. The pipeline let users search inside videos and jump to the exact moment they needed, the kind of feature that sounds simple but unlocked an entire workflow nobody had been able to monetize before.
Recognition
Won the Salesforce AI Pitchfield (2025) with the Trupeer team. Salesforce Ventures awarded $100K.
What I learned
Founding engineer work isn’t about isolated features. It’s about making things work end-to-end, product, infrastructure, performance, reliability, under enough constraint that you have to keep the system simple. Trupeer was a master class in that.