Do you need simple no code machine learning tool?

Harish
6 replies
Yes , I want to simplify and improve productivity of machine learning in short time
No , I will take my own time to build ML models

Replies

Contrarian Maker
For a generic ML modelling. I would never touch a no code tool. On the contrary, if you gave me a no-code image recognition library / tool which I could embed in my product - I would definitely try it. In a nutshell, specific use case >> no code works. Generic use case >> no code sounds scary for ML.
I am founder of ScoopML
@kush_apoorva For now my tool only works for text data which embeded into your product easily
Contrarian Maker
@harishsg99 Got it. It's still 'generic', so a user must decide how and where to use it. Most people who are capable of feeding data into a system (which is a big part of the job) are capable of coding. "No code" revolution is mostly about click-point-solve kinds of things, where the problem is very specifically well defined. I had explored the idea of a no code machine learning tool for *researchers*, who actually can code, but might use something that saves their time. They'd still define all the hyperparameters though. If that's your target market, ProductHunt might not be the best place to recruit early customers.
I am founder of ScoopML
@kush_apoorva thanks for your feedback My tool automatically cleans data and find best possible algorithm for the dataset,it alllows developers to download machine learning model as pickle file.This in fact helps machine learning engineers to improve their productivity by saving time and great accuracy.This tool only works for Regression and Classification problems.
Contrarian Maker
@harishsg99 Aha got it. Sounds pretty cool. When you said no-code I thought you are targeting business people. This is great! This will likely help engineers without ML background be able to implement basic regression and classification jobs. I would not sell it to them as a no-code tool. For devs, I am not sure if it will hit the right nerve. Instead, someone told me that I can run a regression model in minutes with *no boilerplate*, sure as hell I'd try it. A bit like "Add an Artificial Brain to your Product without having to learn ML". All the best.
I am founder of ScoopML
@kush_apoorva thanks for your feedback