Hi all, I'm Louis, i'm the CTO & co-founder at Acoustic Wells , We do IoT/ml solutions for the oil and gas industry. Found Orchest couple days ago on reddit. Since my ML team is mostly pure mathematicians I thought I would give a try to Orchest to try to make their development process easier. Havent tried yet, but i'll give some feedbacks when i'll get everything setup. Thanks for you hard work.

That is awesome to hear! Looking forward to learning more about what we can improve on to make your ML team more productive :)

Welcome Louis! Looking forward to having you evaluate Orchest and know that we're always available in case you need any help 👍

I really like the idea , its very intuitive. Some features I thought about that would make it awesome for us to use: • Build / upload a docker image with the code of all the steps and requirements to an ECR repo • Run the computation on AWS Fargate Question : • lets say california-housing's step Get Data query some data for a given id. - if you have a list of ids for which you'd like to run the pipeline (1 execution per id) how would you do ? Very good job ! Congrats

Hi @Louis thanks for thinking along 💪 Regarding your feature ideas • The code that is run inside the containers is actually not packaged into the containers, it is only mounted. When using custom environment images, you would indeed be able to upload to an ECR repo. Let me know if you need help with this. • We are currently thinking about moving/adding a Kubernetes backend for the containers to run on. On the other hand, having AWS Fargate as the computation backend is not on our roadmap for the near future (but of course this can change). To answer your question: • This is exactly what you can use parametrized jobs for (see the docs). You could add a parameter called

id

with a default value and when creating the job make it a list of ids. The job will then create a pipeline run for every id value in that list.

@Louis I'm Michael, helping with user success at Orchest, just wanted to check in and follow up on this thread. We were wondering if you proceeded with evaluating Orchest. Would you be free to share any updates as to your experience with Orchest (if so)? We're available to arrange a time to talk if helpful. Thanks in advance.