Introduction

Goals

Objective

The goals of this session are to help set you up to use the OpenAI API. We will also look at best practices with respect to workflow for developing and deploying a large application using OpenAI.

Outline

1. Selecting a LLM

There are many to choose from, we discuss what factors to consider

2. Setting up the OpenAI API

Topics include:

• Authentication with key tokens

• Watching usage and costs

• Using the OpenAI Playground to develop prompts

3. Application Workflow

We will go through a case example of extracting author information from a collection of arxiv papers. This is a job that can scale to many thousands of inputs, demonstrates validation of output, and discusses ways to improve reproducibility of results.