Leveraging Machine Learning APIs to Write Better Software
summary
Shows concrete results not hype: a small bookmarklet insta-caption that calls a vision API and replaces missing alt text on Instagram images, a live-caption pipeline for Twitch using a custom speech adaptation trained on about two hours of personal audio, and a measured error around 12.6% on that model. Compares vendors (Google cloud API IBM Watson Microsoft cognitive services), warns about biased training data, and shares tiny, working code examples and GitHub pointers so these fixes actually run.
AI-written summary · about
chapters (11)
- 0:00Start
- 2:52what is machine learning?
- 4:19what are some common uses for machine learning?
- 6:39I'm a lazy programmer
- 7:36IBM Watson Microsoft Cognitive Services
- 8:45am I stuck with the default models?
- 11:45React Redux Redux Saga Express
- 12:23Custom Vision Service Model
- 25:53can we use machine learning?
- 26:25maybe, but we need lots of training data!
- 32:04on a more serious note
related talks
01h 00mBoost your App with Machine Learning APIs
47mBuilding smarter apps with Machine Learning, from magic to reality
27mLive and Machine Learn
17mAI/ML demo derby
49mKeynote 4: Here Come The Robots - Python and Machine Learning
57mBuilding smarter solutions with Machine Learning, from magic to reality
52mBoost Your Apps with Machine Learning APIs
40mMachines are Learning