Mpact scores every applicant across four dimensions, surfaces your top candidates with full score breakdowns, and puts the final decision exactly where it belongs — with you.
Built for African hiring teams. Transparent, configurable, bias-aware.
Publish jobs with skill requirements, experience levels, and education expectations. Candidates apply directly — no login needed. Or import from CSV.
Every candidate is evaluated simultaneously across skills, experience, education, and projects — with configurable weights per role. Done in seconds, not hours.
Ranked candidate cards show score breakdowns, written strengths, gaps, reviewer notes, and bias flags — giving recruiters full context for every decision.
Every score is explainable. Weights are tunable per job. AI reasoning is shown alongside deterministic metrics. Bias flags alert recruiters to potential unconscious bias in the screening — keeping humans in control of every final decision.
Not a black box. Every score has a formula and every recommendation has a reason.
Four-axis formula evaluates skills match via token overlap, experience ratio, education level, and project depth. No randomness — same input always produces the same score.
All candidates are sent to Gemini 1.5 Flash in a single API call. The model reads each profile against the job requirements and produces a holistic fit score, written strengths, gaps, and a recommendation.
Gemini is explicitly instructed to flag name-based, location-based, and institution-prestige bias signals. Any flag is surfaced to the recruiter — not acted on automatically.
Final = Weighted × 0.6 + AI × 0.4. The deterministic base keeps scores consistent; the AI component captures nuance that structured fields miss.
Weights are recruiter-configurable per job. Skills, experience, education, and projects can each be tuned between 0–100%, as long as they sum to 100.
No per-seat fees. No annual contracts. You only pay for what the AI actually uses.
Gemini API calls billed at Google's standard rate. For 100 candidates screened: approx. $0.20 total.
We've watched good people get passed over — not because they weren't qualified, but because their CV landed at the wrong moment, got buried in a hundred-email inbox, or was judged in thirty seconds by someone already fatigued from the previous stack.
Africa's workforce is one of the fastest-growing in the world. The tools supporting it haven't kept up. Enterprise ATS platforms cost thousands of dollars a month and were built for HR teams in London and San Francisco. A startup in Kigali or a growing firm in Lagos shouldn't need a procurement budget just to run a fair, structured hiring process.
Mpact is our answer to that. Structured scoring, written AI reasoning, bias flags, configurable weights — all the things that make hiring both faster and fairer. And built in a way that keeps the recruiter fully in control. The AI surfaces. The human decides.
Mugisha and Principie — two developers from Kigali who got tired of watching brilliant people lose out to slow, opaque hiring processes and decided to build something better.
Builds the systems that make screening fast and honest. Spent too many hours reviewing CVs manually before deciding there had to be a smarter way that still kept humans in the loop.
Built the full stack — from the Flask routes and database models to the recruiter dashboard and applicant-facing UI. Believes software should solve real problems cleanly, without unnecessary complexity getting in the way.
Built during a hackathon. Kigali, Rwanda — 2026.