8 ms·
Show HN: AutoML with API Access
Hi everyone,
I recently built out an automatic machine learning tool to help democratize the use of machine learning. The goal of the project is to help small-medium size businesses incorporate machine learning into their ecosystem without having to spend a lot of time and capital on building out a data science team.
I did some market research and I didn't find any tools that were user friendly or that didn't have a huge cost of entry. The whole idea is that someone who has a base understanding on technology should be able to come into the tool, drop in some data, and then build their own custom experience based on the model created on their data.
I made the user interface as easy as possible to navigate but it still needs some work. Here is a high level summary of the workflow: import your data as a .csv file, choose the column that you want to predict, let Heimdall work its magic, save your model, enable your model with an API key, and voila! You are ready to build your own experience with Heimdall!
Currently Heimdall accepts numerical and single word data but I am currently working on adding in sentence structure support. I have been working on this project as a hobby and using my own funds so the compute is limited. I have successfully built models that have around 30 columns and 20K rows. I understand this may not be good enough for the giant conglomerates but my mission is to help digitize everyday small businesses.
I would love for you to sign up and check out Heimdall! The awesome thing is that none of the original data is being stored anywhere. The only thing that gets stored is the names of the columns and the associated weights. You can always choose to anonymize the names of the columns to add in for additional security.
I am looking forward to any and all feedback. Thank You for your time.
Best,
Joel Reji
- joelreji 3y agoHere is an alpha version demo: https://youtu.be/01gzo6ye4pI https://youtu.be/01gzo6ye4pI