3. From the image to the insight
https://www.capta-astra.com/
In order to address the fragmented software solution landscape in microscopy,
ZEISS took initiative to develop a platform that enables easy software workflow
development and deployment.
Team:
1 Scala – 2 CS
SIP:
Support Unit providing help with prototype to MVP – not all present in product
A light get out dance
Where TL helped
How
Insight not the Image interesting
Cancer or not
Functional: know more about the stuff from other running on our system => max freedom inside the modules restrict on combination
Not starting from zero made it possible for us to come up with a prototype within weeks -> later
DSL – (AST) User defined Workfows
Glue – Marshalling
Runner – start right function with right inputs
In our case containers
Cats:
- All the compiler errors
- Time machine by serialized state : Lookup not in the Runner – side effective modules
Shapeless:
- Marshalling by instanciating case classes to make type verification happen
Monix:
- Fallback
- Parallel
Cool Fact we could compile everything to Scaljs on use it on a Raspberry for a coffe machine cleaning ml workflow.