I'm curious about your ML stack that is also used in production. What has failed, what has given joy?<p>Have you managed to set up a reliable "MLOps" environment with a small(!) team? What are the ingredients?<p>To what extent do you monitor your model inference performance? Is there an automated KPI tracking in place to make sure the new model architecture or a new set of weights perform as expected?<p>How much of your deployment has moved to an "ML Cloud"? Whether it's an AWS, GCP or Azure ML-specific services. Which are the ingredients?
Here's the ML stack I have been using for my last project:<p>- Doing NLP with spaCy (<a href="https://spacy.io/" rel="nofollow">https://spacy.io/</a>) as I consider it to be the most production ready framework for NLP<p>- Annotating datasets with Prodigy (<a href="https://prodi.gy/" rel="nofollow">https://prodi.gy/</a>), a paid tool made by the spaCy team<p>- Deploying the trained spaCy models onto NLP Cloud (<a href="https://nlpcloud.io" rel="nofollow">https://nlpcloud.io</a>), a service I helped creating<p>- Use the models through the NLP Cloud API in production and enrich my Django application out of it