About Us
We built the tool we couldn't find
Candidharbor started as a recruiting problem we kept bumping into, and turned into a company because nobody had solved it the right way yet.
The problem with screening is that everyone knows it's broken, and nobody wants to fix it manually
Sofia spent five years running recruiting ops at two scale-up companies in New York. The problem was always the same: a role would go live, 200 applications would arrive in the first 72 hours, and her team would spend three days triaging before a single interview could happen.
The resumes that got careful attention were the ones at the top of the inbox on a Tuesday morning, not necessarily the strongest candidates. The process was inconsistent by design, because humans are inconsistent by nature.
Max had built backend infrastructure for a workforce management platform. He understood that the resume-to-shortlist step was fundamentally a structured extraction problem: pull signals from unstructured text, compare them against a rubric, rank the outputs. That was a solvable engineering problem.
In early 2025, they started building what they wished had existed. The core design principle: the system explains every decision, or the decision doesn't count. No black boxes, no mystery scores, no "trust the algorithm."
The Team
Three people who've been on both sides of this problem
5 years in recruiting ops at fast-scaling companies before Candidharbor. Led TA function from 30 to 180 hires annually at a logistics platform. Understands what recruiter workflows actually look like at scale.
Built backend infrastructure and data pipelines for HR tech companies for 7 years. Worked on resume parsing and structured extraction before LLMs made it tractable. Designed the scoring architecture from first principles.
Ran implementation and customer success at an HR tech platform for 4 years. Has onboarded more ATS integrations than she can count. Ensures pilot customers run their first shortlist in the first session.
How We Work
What we believe about AI in hiring
A system that produces a score without explaining it isn't a tool, it's a liability. Every Candidharbor score links to evidence. If we can't explain a ranking, we don't surface it.
AI screening reduces the cost of the first pass. It doesn't replace judgment. Candidharbor's job is to give your team better information earlier, not to make decisions for you.
When every candidate is evaluated against the same criteria by the same process, the result is more fair than human screening, which varies by reviewer, time of day, and order of review.
Monthly billing, cancel any time. We're a tool you should keep because it works, not because leaving is complicated. If it stops working for you, you should be able to walk away.