Interview
summary
A one‑person project grows from a homemade conjugation stemmer into a platform with a custom flashcard app, an Alfred‑style lookup, and plans for Android and iOS apps soon. It uses a fine‑tuned open‑source translation model, quantized and optimized for near real‑time subtitle translation, plus a text‑to‑speech engine. Concrete claims: improved translations after many hours of training, live subtitle translations arriving as the caption appears, and a roadmap toward open‑sourcing the language engine.
AI-written summary · about
chapters (7)
- 0:00Introduction
- 1:32Tool to learning platform
- 2:56AI in conjugation and translation
- 4:22Technical challenges
- 6:14Learnings and tooling
- 7:58Future plans and open source
- 10:08Closing remarks
related talks
31mShare information by removing language barriers with Amazon Translate
08mInterview on MBS
39mNative Mobile and AI
21mBoosting Inclusivity: Closed Captioning & Translations in WebRTC
20mAI in Android Development at Duolingo
39mThe Built-in Browser-based Babelfish
21mChalkTalk: Globalizing Data Science Education with AI-generated Videos
23mMaking You a Better Speaker with Spring AI