I’m Rick Sullivan. I’m an engineer at FocusVision, working in my hometown of Portland, OR.
I write about things I’ve recently learned that change the way I work—I’m currently interested in serverless architectures, cloud-hosted machine learning, and natural language processing.
About a year ago, Amazon publicly released Sagemaker, its platform for training and managing machine learning models. Sagemaker has an appealing value proposition–users can train models and tune hyperparameters using on-demand EC2 instances with a variety of hardware options, and models are stored for deployment to endpoints with IAM access control. On my team, we chose Sagemaker for managing our training and deployment of a text classification algorithm using a Tensorflow neural network. Here are some of the things I learned while creating an automated path to production.
In a previous post about Docker for Mac,
I described a way to make SSH keys available inside of Docker containers–here’s
a simpler approach with a few advantages over my previous suggestion. Docker
volumes allow you to link directories from your host machine to the Docker
container, even if the host location is outside of the directory with your
Want to add a contact form to your simple static site? Here’s how to do it with near-zero costs, easy setup, and easy teardown if you want to replace the backend with your own server-based application.
Maybe you’ve seen the slick “Magic Link” login that Slack and other apps now support: here’s how you can support that flow in a Phoenix app.
popd commands in Windows batch or Powershell can be used to move
around the directory stack, like in *nix systems:
I’ve recently gone on a Docker crusade at work.
I’ve found that it’s easy to pass through months at a time, and not be able to answer the question “What have you accomplished recently?”