Terrene: The simplest way to manage machine learning pipelines

Hello Product Hunt 👋 Francois, Cam, and I worked on a few projects together in the past and every time we ended up wasting a whole bunch of time re-building the same infrastructure we had built for previous projects to be able to put our ML models in production. What Terrene does is it abstracts away a lot of the time consuming parts of getting ML into production and allows you to focus on the important parts. With Terrene you can: 1. Create machine learning models with no data science background 2. Log all predictions made by Terrene to assess the performance of the model in real world 3. Schedule automatic re-training if model starts to perform badly 4. Explain predictions made by the model using the predictions exploration tool Here are some of the Terrene's main features: - Model training engine: All you have to do is select variables you are trying to predict and let Terrene take care of feature engineering and model training for you. The trained model will then be accessible through both Terrene's interface and the API. You can checkout the docs on model training to see how simple it really is: https://docs.terrene.co/guides/c.... - Prediction exploration tool: you can use it to examine past predictions and play around with them to try and explain why the model came up with the decisions it did. - Hosted functions: you can extend Terrene with custom python code right inside the interface and do things like sending notifications, pulling in data by web-scraping, etc. - Pipelines: they're hosted cron jobs that you can use to automate tasks on Terrene such as model re-training. - Dashboard: you can use Terrene to create real-time dashboards and visualize the results of the models