For the last few days, I've been exploring a question that came from watching candidates struggle with modern hiringโnot because they lacked ability, but because their experience didn't fit neatly into keyword-based screening.
ATS systems helped companies process applications faster. AI has made resume tailoring even easier.
But it also made me wonder:
Are we helping recruiters understand candidates better, or are we simply getting better at writing for filters?
That question led me to build a small prototype.
Not an ATS.
Not an AI recruiter.
Just an experiment in how AI could evaluate resumes differently.
The idea
Instead of asking:
Does the resume contain the right keywords?
Does the candidate have the exact job title?
Is there a career gap?
The prototype asks:
What behaviours does this role actually require?
What evidence exists in the resume?
Is the evidence specific and credible?
What context should be discussed in an interview rather than assumed?
If this role isn't the right fit, are there other roles that better match the candidate's demonstrated skills?
The emphasis is on explainable reasoning instead of keyword matching.
Current workflow
Understand the role
Read the resume
Evaluate the evidence
Check the context
Generate an explainable verdict
Interview or redirect
Why open source?
I'm not a software engineer.
My background is in customer experience and operations, but over the years I've watched talented people struggle with hiring systems that often reduce complex careers to keywords.
This prototype is my attempt to explore a different approach.
It's intentionally simpleโbuilt with HTML, CSS, and vanilla JavaScriptโto validate the idea before adding more technology.
Looking for feedback
I'd love to hear from:
Recruiters
HR professionals
Engineers
AI builders
Anyone who's experienced modern hiring
If you think the idea has potentialโor you think there are flaws I'd love to discuss them.
GitHub Repository:
https://github.com/devikaliappan1985-code/Evidence-Based-Hiring[]
Top comments (0)