Chatbots in Scala: meet your new robot overlords
summary
Demonstrates a working Scala chatbot that runs CI builds from chat, reduces email and speeds fixes, using an Akka/message-driven architecture and persisting conversation graphs in Neo4j instead of trees. Models conversational state with a monadic conversation model to avoid tangled pattern-matching, and uses Spark ML for training or TensorFlow via gRPC plus cloud NLP (api.ai/Dialogflow) for intent recognition. Notes 99% of platforms use webhooks, recommends heavy mocking and Golden Master tests, and reports the prototype successfully improved team visibility and reaction time.
AI-written summary · about
chapters (24)
- 0:00Introduction
- 1:33Who am I
- 4:04What is a chatbot
- 8:17User experience
- 11:05How Sam got started
- 13:26Why chat bot
- 16:06Chat bot with chart
- 17:55Chatbot toolkit
- 21:56Classic model
- 27:25Data structure
- 31:02Messaging architecture
- 32:10Scala for messaging
- 33:08Natural language processing
- 36:18Machine learning approach
- 37:32Classification models
- 38:12Feedforward neural networks
- 39:49Machine learning vs deep learning
- 40:14Scala API
- 40:46Tensorflow
- 41:19Why not external services
- 43:09Testing Chatbots
- 44:09Building your own Chatbot
- 44:43Reference User
- 44:59Another story
related talks
13mExploring AI-Powered Chatbots and Their Transformative Role
54mBuilding advanced ChatBots
55mEnterprise Chatbots (using messaging and AI to build novel user interfaces for work)
27mBuilding AI Agents with Spring & MCP
01h 02mBuilding full stack AI chatbots with Java - livecoding
19mThe Future of AI
30mConversational AI: Building clever chatbots
47mTesting Conversational AI