Skip to main content
강홍재/ James
← Work
MVP2026 · Founder · Product · QA Engineer· Started(First Commit date)

ReleaseGate

An AI evaluation SaaS that turns release-or-hold decisions into data - just before launch.

  • SaaS
  • AI
  • FastAPI
  • React
Problem

In release meetings, "can we ship this?" is decided by gut feel and whoever speaks loudest. There's no place where QA, PM, and engineering can look at the same number and agree.

Context

Fourteen years in QA showed me the same pattern in release meetings, repeatedly: long checklists, no single answer for "release confidence," so the call falls to whoever sounds most certain.

Users

PO/PM holding the release call, QA leads watching the risk, engineering managers locking the date.

Hypothesis

Express release readiness as a single 0-100 confidence score plus a GO / HOLD recommendation, and the conversation shifts from debate to verification.

What I did
  • Designed the Release Confidence Score (0-100) - weighted blend of coverage, open defects, and change-risk areas
  • Report UI that shows the GO / HOLD recommendation alongside its reasoning
  • AI-drafted test case starters so QA opens an editable draft rather than a blank screen
  • PDF report export for stakeholder sharing
Product decisions
  • Single headline number on the score - so a meeting can begin with "what's the number?" instead of debate
  • GO/HOLD shown as a recommendation only, never a verdict - acknowledging the limits of automation
  • PDF export over integrations for the MVP - minimizing adoption friction
Metrics

MVP stage. Early user interviews in progress; quantitative metrics not yet measured.

Result / Learning

Confirmed how hard it is to fit "the weight of a release decision" onto one screen. The next round tests the hypothesis that adoption depends less on the number itself and more on how the reasoning behind it is presented.

Outlook

The project being pushed for commercialization most actively. Landing the first paying team is the next step.

QA lens on this call

The starting point came from years of QA engineers saying "all of this is in the tracker - why don't decision-makers see it?" Information existing and information shaped for a decision are two different problems; that insight set the feature priorities.

Tech stack
  • FastAPI
  • React
  • PostgreSQL
  • OpenAI API