Inspiration

Coding agents can generate implementation quickly, but engineering teams still need trust: what changed, what is risky, what tests prove readiness, and who approved the release decision.

What it does

Enterprise AI Engineering Orchestrator gives engineering and product teams a compact release-readiness workflow: run a risk scan, review affected areas, generate test actions, collect evidence, and make a visible human approval decision.

How we built it

The MVP is a React, TypeScript, and Vite application with local simulated workflow state. It is structured around panels for release risk, repo context, agent workflow, impacted tests, human approval, and evidence. The same codebase can be positioned for UiPath, GitLab, and product-focused hackathon submissions.

Challenges

The main challenge was avoiding a generic chatbot demo. The useful product surface is a workflow: context, risk, generated actions, evidence, and human approval.

Accomplishments

  • Built a responsive dashboard MVP.
  • Added per-hackathon positioning modes.
  • Implemented scan, generated-test, and approval states.
  • Prepared CI, README, MIT license, and demo script.

What's next

Next steps are product analytics instrumentation, real GitLab context, UiPath Automation Cloud approval routing, and real generated test patches.

GitHub repo: https://github.com/zemeng2015/enterprise-ai-engineering-orchestrator

Public proof package

Novus / Pendo proof

Novus/Pendo is installed on the deployed GitHub Pages demo. Public dashboard proof:

https://raw.githubusercontent.com/zemeng2015/enterprise-ai-engineering-orchestrator/main/docs/novus-dashboard-live-after-install.png

The live demo also includes a Submission Proofs panel and machine-readable proof index: https://zemeng2015.github.io/enterprise-ai-engineering-orchestrator/submission-proof-index.json

Built With

  • ai-agents
  • product-analytics
  • qa-automation
  • react
  • release-management
  • typescript
  • vite
Share this project:

Updates