vaibhavdubey97

5 ms·

1. 1.▲
   by vaibhavdubey97 11mo ago

   Thanks a lot! We’ve tried to mimic the workflow of an ML engineer and built agents that can own specific functions of the workflow. Good to hear that the idea resonates with you!
2. 2.▲
   by vaibhavdubey97 11mo ago

   Makes sense! We'll be sure to make this available very soon :) Thank you!
3. 3.▲
   by vaibhavdubey97 11mo ago

   Thanks for your feedback! The "export analysis" functionality is built to enable you to get detailed data insights and get a report generated from the insights. Would you prefer to see the entire chat in your export or would it be
4. 4.▲
   by vaibhavdubey97 11mo ago

   Absolutely! Getting AI agents to determine the right set of features from raw data has been a very interesting problem
5. 5.▲
   by vaibhavdubey97 11mo ago

   Thanks a lot! On a side note: big fan of mljar here. When we were initially playing around with using agents for automating ML tasks, we had used problems from the openml's automl benchmark which you had posted about on Reddit for our
6. 6.▲
   by vaibhavdubey97 11mo ago

   Thanks a lot! Excited for you to try it out and get your feedback :)
7. 7.▲
   by vaibhavdubey97 11mo ago

   Thanks for the great feedback! We've added a `baseline_deployed` status where the agents create an initial baseline and deploy it so you have something to play around with quickly. This is why you're seeing a blank json there. Onc
8. 8.▲
   by vaibhavdubey97 11mo ago

   We have a data enricher feature (still in a beta mode) which uses LLMs to generate labels for your data. For cleaning and feature engineering, we use agents that automatically handle it for you once you've connected your data and defin
9. 9.▲
   by vaibhavdubey97 11mo ago

   Thank you! :)
10. 10.▲
    Launch HN: Plexe (YC X25) – Build production-grade ML models from prompts(plexe.ai)

    85 points by vaibhavdubey97 11mo ago|31 comments
11. 11.▲
    by vaibhavdubey97 1y ago

    Absolutely! We started off thinking that we wanted to automate the whole way in one go and then add restrictions and interruptions based on areas where users face issues. This is great feedback so thank you!
12. 12.▲
    by vaibhavdubey97 1y ago

    Hey! That sounds great. Happy to help in case you face any issues while adding support for vertex.ai. We’ve added a tool which does EDA and when the model package is created, it contains a file called metadata.json which has detailed explan
13. 13.▲
    by vaibhavdubey97 1y ago

    Thank you! Making sure that basics don't going anywhere :)
14. 14.▲
    by vaibhavdubey97 1y ago

    Agree completely. We built Plexe with that first scenario in mind - the messy spreadsheet problem that's so common in enterprise. You can connect multiple data sources, and Plexe will identify what it needs based on the problem descrip
15. 15.▲
    by vaibhavdubey97 1y ago

    Thank you! Hopefully more complicated transformers are coming soon too :)
16. 16.▲
    by vaibhavdubey97 1y ago

    Absolutely! When we started building this out, we knew that we had to build an agent to perform data cleaning and feature transformations. After speaking to data analysts, PMs and engineers over the last few weeks, we've received stron
17. 17.▲
    by vaibhavdubey97 1y ago

    phew that was close. I'm glad your faith is restored :)
18. 18.▲
    by vaibhavdubey97 1y ago

    Sorry I think I explained poorly. Plexe does build deep learning models automatically. When it gets a dataset and a problem description, it automatically evaluates various model architectures (NNs being one of them). Plexe experiments with
19. 19.▲
    by vaibhavdubey97 1y ago

    Plexe analyzes your data and task description, then builds custom ML models using standard Python libraries (like scikit-learn, XGBoost, etc.). If your problem is best solved by a regression model, it will build that. If classification is m
20. 20.▲
    by vaibhavdubey97 1y ago

    Very cool, thanks for sharing! :)
21. 21.▲
    by vaibhavdubey97 1y ago

    Absolutely! And hopefully an input/output schema for the model :)
22. 22.▲
    by vaibhavdubey97 1y ago

    Thank you! Smolagents works great for us but we did run into some limitations. For example, it lacks structured output enforcement, parallel execution, and in-built shared memory, which are crucial features for orchestrating a multi-layer a
23. 23.▲
    by vaibhavdubey97 1y ago

    You're right. We've seen the "garbage in, garbage out" problem firsthand. We've seen the models hit typical statistical pitfalls like overfitting and data leakage during testing. We've improved by implementing
24. 24.▲
    by vaibhavdubey97 1y ago

    Thank you! Yeah we rushed to create a "Plexe in action" video for our Readme. We'll put a link to the YouTube video on the Readme so it's easier. Using large generative models enables fast prototyping, but runs into seve
25. 25.▲
    Show HN: Plexe – ML Models from a Prompt(github.com)

    130 points by vaibhavdubey97 1y ago|49 comments