Back to browse
Decision Audit – A deterministic decision engine for career pivots

Decision Audit – A deterministic decision engine for career pivots

by cosmok·Feb 15, 2026·2 points·1 comment

AI Analysis

Mid

Avoids LLM hallucination with deterministic scoring, but a pros/cons spreadsheet solves the same problem.

Strengths
  • Smart framing: rejects chat-based AI for structured, weighted decision logic—philosophically sound.
  • Multi-faceted output (Burnout, Runway, Decision Rights indices) goes beyond generic advice.
  • Clean, focused UX for a high-stakes decision—not overwhelming.
Weaknesses
  • Core problem (pros/cons for decisions) already solved by spreadsheets, decision matrices, career coaches.
  • No evidence the indices are validated or better than existing frameworks; proprietary scoring without external research backing.
Category
Target Audience

Career-uncertain professionals evaluating job transitions

Similar To

spreadsheet templates · career coaching frameworks · decision matrix tools

Post Description

Hello HN,

I’m a solo dev building a tool to replace the "pros and cons" list for high-stakes decisions.

While LLMs are great for brainstorming, I found them dangerous for actual decision-making—they hallucinate and try to please you rather than giving you a hard truth. I wanted something that combined deterministic logic (weighted criteria) with generative explanation (narrative).

The Product: It’s called Decision Audit. The first module is "Should I Quit?" (career transition analysis). Instead of a chat interface, it uses a structured interview to calculate specific indices (Burnout, Runway, Decision Rights, Optionality). The output isn't a text blob; it's a 3-page structured Dossier (PDF) that includes:

An "Exit Lean" score (calculated via weighted algorithm, not LLM sentiment). A "Risk Matrix" (flagging specific dangers like visa dependency or low runway). A 30-day "Validation Sprint" (action plan).

The Stack: Frontend: Next.js (focused on mobile-first, fast forms). Backend: Node.

The "Brain": A rules-based scoring engine (JSON logic) handles the math. I only use the LLM API at the very end to summarize the data points into the narrative sections of the PDF.

Business Model: I’m bootstrapping this and experimenting with a "Pay What It's Worth" model for launch. No accounts required to take the audit. No tracking pixels or ad networks or storing of submitted information.

I’d love feedback on the scoring weighting—specifically if the "Chaos Index" feels too sensitive.

Link: https://decide.trk7.app/intake/should-i-quit

Similar Projects