Quality assurance · Production

Duolingo boosts developer speed up to 25% with GitHub Copilot and Codespaces

The problem

Duolingo's developers needed to be as efficient as possible, but fragmented tooling across repositories—including third-party tools like Gerrit and PullApprove—created inconsistent workflows and prevented developers from moving easily between projects.

First attempt

Relying on third-party tools like Gerrit and PullApprove for code review left Duolingo's primary repositories with widely varying cultures and pull request processes, creating inefficiency and preventing developers from moving easily between repos.

Workflow diagram · grounded in source
1
Developer initiates coding task
trigger
“by starting to write the code they want to use or by writing natural language comments that describe what they want the code to do”
2
Copilot autocomplete suggestions
ai_action
“an AI-powered pair programmer that provides autocomplete-style suggestions to developers while they code”
3
Context-aware recommendations
ai_action
“GitHub Copilot is unique in the sense that it looks at the context of the rest of your work and incorporates that context into its recommendations”
4
Boilerplate tab completion
ai_action
“Boilerplate code is where Copilot is very, very effective. You can practically tab complete the basic class or function using Copilot”
5
Codespaces environment setup
integration
“Codespaces offered a way to skip the local environment troubleshooting and offer a 1-click environment setup”
6
Slack code review integration
integration
“one Slack integration has dropped the median turnaround time for code review from three hours to one”
Reported outcome

GitHub Copilot increased developer speed by at least 25% for those new to a codebase and 10% for familiar ones; a custom Slack integration cut code review turnaround from three hours to one; and Codespaces reduced the largest repo setup time from hours or days to one minute.

Reported metrics
Developer speed increase (new to codebase)at least 25%
Developer speed increase (familiar with codebase)10%
Code review median turnaround timefrom three hours to one
largest repo setup time after Codespacesone minute
Show all 6 reported metrics
developer speed increase (new to codebase)at least 25%
developer speed increase (familiar with codebase)10%
code review median turnaround timefrom three hours to one
largest repo setup time after Codespacesone minute
developer productivityincreased developer productivity
routine work timespend less time on routine work
Reported stack
GitHub EnterpriseGitHub CopilotCodespacesGitHub's APIsGerritPullApproveSlack
Source
https://github.com/customer-stories/duolingo
Read source ↗

Frequently asked questions

What did this team achieve with this AI workflow?

GitHub Copilot increased developer speed by at least 25% for those new to a codebase and 10% for familiar ones; a custom Slack integration cut code review turnaround from three hours to one; and Codespaces reduced the…

What tools did this team use?

GitHub Enterprise, GitHub Copilot, Codespaces, GitHub's APIs, Gerrit, PullApprove, Slack.

What results were reported?

Developer speed increase (new to codebase): at least 25%; Developer speed increase (familiar with codebase): 10%; Code review median turnaround time: from three hours to one; largest repo setup time after Codespaces: one minute (source-reported, not independently verified).

What failed first in this deployment?

Relying on third-party tools like Gerrit and PullApprove for code review left Duolingo's primary repositories with widely varying cultures and pull request processes, creating inefficiency and preventing developers fr…

How is this quality assurance AI workflow structured?

Developer initiates coding task → Copilot autocomplete suggestions → Context-aware recommendations → Boilerplate tab completion → Codespaces environment setup → Slack code review integration.

Related quality assurance cases