Internal AI developer tooling

Dartly

Designed a GPT-powered internal tool that converts business-specific JavaScript patterns into Dart to accelerate Flutter development.

Role

Product engineer

Team

Internal initiative

Timeline

2024

Outcome

50% less manual conversion work

Problem

Teams were spending too much time manually translating business-specific JavaScript logic into Dart during product development.

Role

Designed and developed the internal AI conversion workflow around business-specific code patterns.

Approach

  • Mapped repeatable JavaScript translation patterns that were slowing down implementation.
  • Built a custom GPT-driven conversion workflow tailored to business rules rather than generic code translation.
  • Focused on responsible output quality by narrowing the tool to high-value use cases the team could validate quickly.

Impact

  • Cut code conversion effort by half for common workflows.
  • Improved development velocity on Flutter features that depended on pre-existing JavaScript business logic.
  • Demonstrated how targeted AI tooling can solve narrow but expensive engineering bottlenecks.

Key learning

AI tools create the most value when they are scoped to a narrow operational problem with clear validation rules and fast human review loops.

Next step

If this case study is relevant, the next move is straightforward.

You can review the broader portfolio, move to the next case study, or contact me directly for a conversation around similar product and engineering work.