Watson and the Cognitive Developer
summary
A live Enterprise Java candidate-adviser called Esaki runs on Bluemix, calling Alchemy Language (custom model from Watson Knowledge Studio), Conversation, Document Conversion and Cloudant; resumes go through Document Conversion then Alchemy for entity extraction, TF‑IDF and custom-dictionary scoring produce ranked job matches. The pilot is live on IBM.com careers, collects interactions for weekly retraining and JUnit tests, and is production-ready but still misclassifies some intents during early three-week usage while the team tweaks annotations and context handling.
AI-written summary · about
chapters (11)
- 0:00Start
- 9:46Architecture
- 12:19Use Annotations to enhance response
- 15:01Context points to Candidate Object
- 15:54Off Topic intent is a reality in cognitive apps
- 16:46Calculating scores to match jobs & job categories
- 20:43Enriched Candidate object
- 21:48Enriched Jobs Object
- 22:26Calling AlchemyAPI....
- 22:57Custom Entities in Watson Knowledge Studio
- 23:28Pre-Annotation with Custom Dictionaries
related talks
07mInterview with Jonathan Kaufman
14mArtificial Intelligence in Services
19mKeynote IBM: "From old to new IBM, leading to cognitive era"
13mDiscusses the "Candy Store" of IBM Emerging Technologies
27mHow to put AI to work in different industries
38mBuilding Smarter Applications Using Cognitive Computing
48mScale and accelerate the impact of generative AI with watsonx
16mOutthink machines coping with humans