I am working on a new product, and I cannot decide on a good name.
I originally wanted to use ‘ML Studio’, but then I found out that there are multiple tools out there using the same name. So I tried coming up with a few other names, but none really ‘sticks’. What do you think I should pick?
- ML Creator
- ML Studio
- Data Studio
- Something else
So what is it? What are we naming?
I created a partitioned application that is deployed simply as packages in OpenFlow, and can run on any of the three agent types.
- A frontend that allows you to create machine learning models in a visual designer Including both layered and graph models, as well as pre-built models downloadable from Kaggle. It also allows you to create workflows that enable getting, transforming, and sending data into the machine learning models (both training and predicting).
- A backend that hosts the running workflows and models.
Splitting this up allows hosting the frontend inside OpenFlow, while hosting the models on dedicated hardware. It also makes it possible to scale the platform more efficiently.
Why?
Well, OpenFlow came out of the dream to create:
- A platform that allows for the storage of any type of data in a secure and scalable way.
- A platform that facilitates the easy, flexible, and secure collection of events and data.
- A platform that enables flexible responses to streams of data and events.
- A platform that simplifies reporting and enhances learning about your data and events.
- A platform that allows enriching data using code and AI in an easy and flexible way.
Working on collecting data took me far around different technologies like RPA (OpenRPA) and human workflows (BPM and Forms) and IIoT (talking to SCADA, PLCs using mod/profibus, OPC, S7, iec/60870-5 and iec61850, and working inside OT network, Lora, NB-IoT, and more).
Adding reporting took longer than anticipated. However, it was important to me that it wasn’t just something I had created, but rather something that extended the authentication and Access Control Lists from OpenFlow into the platform.
This leaves the addition of AI pipelines. I should have begun this aspect a long time ago. However, it was during my work on LLM integration that I gathered the courage to embark on this crucial part – not the final one, but ‘final’ in terms of my initial vision for the platform seven years ago.