TalentGrid

Building Smarter Hiring for New Zealand

The New Zealand job market has unique dynamics — tight talent pools, a strong SME base, and hiring managers wearing many hats. Generic job board tools weren't cutting it.

We started TalentGrid with a discovery sprint to map recruiter pain points: too many irrelevant applications, no easy way to track candidate progress, and back-and-forth email chains eating the day.

The Matching Engine

The core of TalentGrid is a Python-based matching service that extracts structured data from CVs using a combination of rule-based parsing and LLM-assisted extraction. Each candidate receives a scored fit profile against the job's required competencies, experience level, and location preferences.

The score isn't a black box — recruiters see a breakdown of why a candidate ranks where they do, building trust in the system.

Pipeline and Collaboration

The hiring pipeline is a Kanban-style board backed by a PostgreSQL state machine. Role-based access lets hiring managers leave feedback, scheduling links auto-send at stage transitions, and email/SMS notifications keep candidates in the loop without manual follow-up.

Infrastructure

The platform runs on AWS with a Next.js frontend deployed via Vercel, a FastAPI backend on ECS, and a Neon PostgreSQL database. The architecture was designed to scale to multiple recruiting agencies from day one.