Last month a small startup during IBM Watson Warsaw Summit presented their robotics solution empowered by IBM Watson Visual Recognition. The audience asked the startup’s CFO why they didn’t invest to build their own AI capability. For CFO the bottom line seemed to be crucial at the end of the day, and with the power, and the price IBM Watson cognitive service available from IBM Cloud made a lot of the sense. But the question started to bother me…

As a developer advocate I have been helping a lot of startups with Watson cognitive services. They developed various technologies alike the one for the robot security solution based on the Watson Visual Recognition to trace people and objects.

I have been thinking about the challenge of building an AI system from scratch. But when today I learned about IBM announcing the partnership with Lightbend everything found its place.

IBM & Lightbend announced the collaboration on developing the Reactive Platform business

IBM & Lightbend will now be collaborating on developing the Reactive Platform business. Starting with the existing Watson ML integration into DSx, IBM and Lightbend will enable the deployment of the Watson ML models into the Reactive Platform development & production environment. This will leverage the Scala Build Tool, and also rely on open source extensions to Spark that will be built by the Spark Technology Center. Spark is the natural conduit for integration due to the fact that DSx is built on Spark and it is deeply embedded into the Reactive Platform as well.

In other words I have found Scala — the framework to develop, define, and train the AI systems complimentary to Watson cognitive services. So based on the various cognitive services I can create the hybrid AI system with my own secret sauce and external commodity cognitive services.

Subscribe to my blog to see the next steps in using Scala with IBM Watson on IBM Cloud with robots and beyond.


1. IBM Announcement on Lightbend:

2. Massive Online Open Course on robots:

3. Check this GitHub:

Subscribe, and follow me on Twitter: Marek Sadowski