Elite Strategist

Inspiration

Most resume tools optimize keywords or formatting, but hiring decisions are shaped by human judgment, ATS filters, and tradeoffs across roles and markets. I wanted to build something that reveals how decisions are actually made and how small wording changes or career choices can meaningfully change outcomes. The idea was to turn resume feedback into a simulation of the hiring ecosystem, not just a checker.

What it does

AI-strategist is a multi-perspective hiring decision simulator. It analyzes a resume and:

Simulates recruiter personas (Big Tech recruiter, startup founder, hiring manager, ATS)
Performs semantic JD ↔ resume alignment (meaning-level gaps, not keyword matching)
Runs counterfactual “what-if” career paths (how alternate choices compound over time)
Audits claim credibility and suggests defensible rewrites
Generates a 90-day action plan to close gaps for target roles and regions

Instead of telling users what to change, the app shows how different decision-makers interpret the same profile and how changes shift outcomes.

How we built it

The app uses Gemini 3 as the core reasoning engine to:

Reason over entire resumes and job descriptions (long context)
Simulate multiple stakeholder perspectives with structured outputs
Perform counterfactual scenario modeling for alternate career paths
Produce explainable rewrites and action plans

The frontend provides a guided workflow (resume input → target roles → personas → scenarios), while the backend orchestrates Gemini 3 calls and renders structured results into interactive panels (Recruiter POV, JD Gap Map, Counterfactuals, Claim Verification).

Challenges

Avoiding generic advice: We constrained outputs to be defensible with recruiter rationale and concrete rewrites.
Explainability: Making multi-persona reasoning understandable required structured JSON outputs and careful UI design.
Scope control: Balancing depth (simulation, counterfactuals) with demo clarity within a hackathon timeline.

What we learned

Hiring outcomes depend on interpretation, not just credentials.
Small phrasing changes can meaningfully alter shortlist likelihood across personas.
Gemini 3’s long-context and multi-perspective reasoning enables decision simulation, not just text generation.