Aquarium’s mission is to make it easier to build and improve production ML systems. Our team has a lot of experience building and deploying ML products and we want to make it easier for everyone to solve practical problems with smart artificial intelligence.

Aquarium is hiring for a contractor Site Reliability / Platform Engineer role, focused on our Tidepool product. Tidepool is a data enrichment application that sits on top of a data warehouse, and allows users to create derived AI tables based on existing text columns. To address security and privacy needs from larger target customers, we are adapting Tidepool for single-tenancy and on-prem / customer-hosted deployment models. 

In this role, you will provide expertise and engineering to move us away from a SaaS multi-tenant deployment model tied to GCP, to a more vendor-agnostic Kubernetes structure. We have already been working on the application logic and architecture changes to do that decoupling, and so this role focuses around topics like service configuration, packaging, logging, etc. Our core stack consists of Python, Postgres, Redis, Typescript, and Celery.

We are flexible in specific hour commitments, but would prefer a minimum commitment of 15+ hours per week, and are open to discussing full-time contracting as well.

Responsibilities

• Set up single tenant deployments of our stack across multiple cloud environments
• Iterate on local developer tooling to match the updated infrastructure and configuration
• Configure a network architecture that’s compatible with both multi-tenant SaaS and single tenant deployments.
• Converting terraform and infra management scripts into a unified deploy architecture.
• Establish best practices for configuration, monitoring, logging, and release for on-prem deployments.

Qualifications

• Strong experience in Kubernetes, including both initial service setup and production maintenance
• Strong experience with single-tenancy, customer-cloud, and other “on-prem” deployments
• Familiarity with production Redis management
• (Plus) Familiarity with columnar data warehouses such as BigQuery, Redshift, or Snowflake.

Our current focus is developing a product called Tidepool. Tidepool is a data analytics tool for large unstructured text datasets.

Tidepool lets you categorize and surface patterns in text so you can derive valuable insights with your existing business intelligence workflow. Answer questions like, “what topics tend to have the highest engagement?" or "what categories of businesses are mentioned in these bills?" over text datasets of 100M+ rows.