Machine Learning Deployed in the Real-World

Release:

3.0.0

Date:

Feb 08, 2023

---

Sermos provides a set of tools and design roadmap for developers to quickly and effectively integrate Data Science into real-world applications. The core design decisions and recommended technologies are based on nearly a decade of putting data science to work in demanding applications such as real-time motorsports strategy at Pit Rho, custom implementations across industries as diverse as energy, finance, and healthcare at Rho AI, and climate impact assesment tools at CRANE. With that said, this is built by developers and data scientists and strives to strike an appropriate balance between opinionated decisions and personal choice.

This is open source software and we look forward to seeing what you can build on top of Sermos. For those looking for quick, scalable, enterprise-grade deployments, make sure to check out Sermos Cloud, which is purpose-built for running complex, scalable, highly available Machine Learning (ML) workloads, including Natural Language Processing (NLP), Computer Vision (CV), Decision Modeling, Internet of Things (IoT), and more.

Sermos comprises a few key libraries:

1. Sermos

• This is the base package with optional scaffolding, architecture components (e.g. DAG implementation on top of Celery), and some useful utilities.

1. Sermos Tools

• Tool catalog for use in Sermos (or other!) Machine Learning applications.

To start using Sermos (see: Quick Start Guide).

For anyone using Sermos Cloud, you need only have:

1. A Python package with a properly configured setup.py

2. A simple sermos.yaml configuration file.

3. Sermos handles the rest.

Contents

Core

• Quick Start Guide
  • Get running in 5 minutes or less!
• Overview
• Deployments
  • Prerequisites
  • Deployment
  • Deployment Status
  • Proper Python Package Structure
  • Sermos Deploy
• Sermos Yaml
  • Yaml Format
• CLI Tools
  • Deployments
  • Proxy

Supporting Libraries

• Sermos Tools
• Rho ML

• Index

• Module Index