For most companies, the hiring process begins the same way: a recruiter opens a folder of resumes and starts reading. Across a pool of 100 candidates, that manual first-pass screening can consume more than 80 hours of interviewer time before a single meaningful conversation takes place. Singapore-based AI company Lumina believes that the number should be closer to zero.

Lumina recently launched Aria, an AI hiring agent designed to overhaul first-pass candidate screening. The product conducts structured, voice-based interviews asynchronously, evaluates candidates across three dimensions, and delivers ranked outputs that hiring teams can review in under a minute per candidate — a reduction the company estimates at up to 50 times faster than current manual processes.

At the heart of Lumina’s proposition is what Glenn Low, the company’s CEO and co-founder, describes as a signal-to-noise problem in hiring.

“The signal is whether a candidate can actually do the job,” Low explains. “The noise is everything that obscures that signal — inflated resumes, keyword optimisation, and subjective interpretation during early screening.”

As AI-assisted resume writing becomes more widespread, Low argues the problem compounds. Candidates optimise their documents for keywords. Recruiters apply shortcuts — university names, previous employers, gut feel — that introduce variability into what should be a consistent evaluation process. Resumes, as Low puts it, are “2D representations of 3D people,” inherently incomplete documents that companies compensate for with first-round interviews.

Also Read: From HR to talent flow: Why workforce management needs a supply chain mindset

Aria is Lumina’s attempt to restructure that early funnel before human judgment enters the picture.

How Aria works

The process begins with resume-to-job description matching, which shortlists candidates before any interview takes place. Those who advance receive a link to complete a voice interview at their own convenience, eliminating the need for calendar coordination between recruiters and candidates.

During the interview, Aria follows a structured flow designed to assess critical thinking, domain knowledge, leadership, and collaboration. The system uses adaptive probing based on candidate responses rather than following a rigid script, evaluating answers for quality, depth, and relevance. A sentiment analysis layer runs in parallel, assessing tone and authenticity across the conversation.

The outputs are consolidated into scores across three areas: resume-to-job description match, interview performance, and sentiment. Crucially, Lumina has built explainability into the scoring. Each result comes with supporting insights showing how the score was reached, giving hiring teams a clear rationale rather than an opaque number.

AI hiring tools have attracted meaningful scrutiny over the past several years, with critics pointing to the risk of encoding existing biases into automated systems. Lumina’s response centres on the consistency of its scoring rubric. Because Aria evaluates every candidate against the same defined criteria, the company argues it removes the subjective interpretation that drives variability in human-led screening.

Also Read: AI is removing the co-founder bottleneck for early-stage startups

Whether that fully addresses bias concerns — particularly around how rubrics are designed and validated — is a question the industry is still working through. Low acknowledges the responsibility, noting that the team has deliberately built explainability into every score to ensure hiring teams can interrogate the system’s recommendations rather than simply accepting them.

Where humans stay in the loop

Lumina is deliberate about where Aria stops. The product is positioned as a first-pass tool, not a replacement for human judgment across the full hiring process. Once Aria surfaces its ranked shortlist, recruiters take over to determine who advances to later-stage interviews.

“As candidates progress further into the process, human judgment becomes increasingly important,” Low says. “Later-stage interviews focus on nuance, culture fit, and building relationships — areas where human interaction adds the most value.”

Beyond enterprise hiring teams, Lumina has also seen demand from job seekers directly. The company’s resume analyser allows candidates to understand how their profiles are likely to be interpreted — a move that Low describes as shifting from guesswork to a more structured view of their own experience.

Lumina’s focus for now remains squarely on first-pass screening, the stage where hiring volume is highest and inefficiency most acute. Whether Aria eventually moves deeper into the funnel will likely depend on how well it earns trust at the stage it has already chosen to own.

—

Image Credit: Lumina

The post Lumina’s Aria aims to fix what is broken at the top of the hiring funnel appeared first on e27.