Thread by @chipro, Some resources that I’ve found really helpful to understand machine learning in [...]

Some resources that I’ve found really helpful to understand machine learning in production.

1. Engineering starts with infrastructure. @vtuulos gave a great overview of the relationship between data science and infrastructure at Netflix

2. What and how to monitor ML systems in the wild. @josh_wills gave an excellent deep-dive on DevOps meets Data Science based on his experience at Google, Cloudera, and Slack https://www.infoq.com/presentations/instrumentation-observability-monitoring-ml/

Instrumentation, Observability & Monitoring of Machine Learning Models

Josh Wills discusses the monitoring and visibility needs of machine learning models in order to bridge gaps between ML practitioners and DevOps.

https://www.infoq.com/presentations/instrumentation-observability-monitoring-ml/

3. Deploying ML is easy. Deploying it reliably is hard. Daniel Papasian and @tmu analyzed post mortems of 96 ML systems outages at Google and found that most outages are even ML and more related to the distributed character of the pipeline

4. I can’t believe I’m sharing a VC post, but @martin_casado and @BornsteinMatt gave an interesting perspective on the economics of AI, how cloud services are reducing the margin, scaling problem due to edge cases, and the diminishing value of added data https://a16z.com/2020/02/16/the-new-business-of-ai-and-how-its-different-from-traditional-software/

The New Business of AI (and How It's Different From Traditional Software)

We have noticed in many cases that AI companies simply don’t have the same economic construction as software businesses. At times, they can even look more like traditional services companies.

https://a16z.com/2020/02/16/the-new-business-of-ai-and-how-its-different-from-traditional-software/

You can follow @chipro.

Tip: mention @twtextapp on a Twitter thread with the keyword “unroll” to get a link to it.

Latest Threads Unrolled: