There is one uncomfortable truth: the core signals in hiring are self‑reported. A CV is what someone says about themselves. A portfolio is what they choose to show and how they choose to frame it. Even references are often hand‑picked advocates. In an era of coding AIs, this becomes an adversarial game: candidates are optimizing for getting through filters and interviews, “how can I optimize my CV better so it reaches a human to take a look at it instead of getting instantly rejected because it did not fit an imaginary frame from that company’s filter”, not for accurately representing their abilities.

AI has supercharged the problem. It is now trivial to generate tailored resumes, cover letters, even semi‑plausible code samples and project write‑ups. On the other side, companies lean on AI‑driven screening to cope with volume. The result is an arms race: smarter keyword stuffing versus smarter filters. What does not improve in that battle is your confidence that the person you hire can actually do the work. And for job seekers, it is spending more time on improving your CV great and portfolio instead of your skills and abilities to learn something new.

Recruiters systematically favor people who know how to play the credential game—and quietly filter out people who built their skills in less conventional ways: smaller markets, scrappy startups, open‑source communities, self‑directed learning.

Even “skills‑based hiring,” as most companies practice it, is theatre. The language changes, but the inputs are still weak: self‑reported skills sections, generic online tests, puzzle‑style coding challenges that correlate more with test prep than with real‑world engineering. Most of the self-taught people won’t be able to know to answer these interview puzzles with lack luster questions that are given during interviews in the hiring process, since they are not part of the job or worse, the best interviewers rather than best doers will instantly answer them, and be considered an amazing candidate and hired, which then leads to a bad hire in the company that leads to waste of time, money, effort and especially is bad for the moral of the team. There is nothing that kills morale in a team, than having a bad hire and an under performer who was presented as great candidate.

Leaders tell themselves they are hiring for skills, but in practice they are still hiring for signal fluency—who speaks the right jargon, who has the right logos, who can navigate a contrived interview loop.

Meanwhile, the cost of getting it wrong is high. A bad technical hire does not just waste a salary line; it drags down an entire team, blows up roadmaps, and forces strong people to spend their time cleaning up after weak ones. The harder and stressful the work, the more that gap shows. And because the inputs were never truly trustworthy, post‑mortems often end with a shrug: “They looked great on paper.”

Some companies have found some way to mitigate this risk and avoid bad hires, one of the most important thing is hiring for personality rather than solely for the skills. Skill can always be taught, but personality cannot be.

Underneath all the surface issues—ATS tuning, sourcing channels, employer branding—sits the real failure mode: hiring decisions are being made on claims and impressions, not on trusted, transferable proof of skill.

The Solution: Build a Skills Trust Layer with Peer‑Verified Proof Fixing this is not about one more tool in the stack. It is about changing what your organization treats as truth in hiring.

Right now the truth comes from two places: • What the candidate says (CV, portfolio, the interview narrative) • Company own research, which can be hard to obtain and costly if done for every candidate or if the candidate is a private person for any type of reasons

This foundation is fragile and costly, and it has proven that it does not bring much of results, it is especially bad for start ups that want to bring quality talent, but don’t have the resources at this stage to do fully detailed checks who they are hiring.

A more robust foundation is peer‑verified skills: moving from “I claim this”, “I can do this”, “Oh yeah I don this” to “Google engineer had verified I can do this, check it out here is proof”

In practice, that means treating each key skill as a concrete, assessable claim tied to evidence: • Instead of “React” on a CV, you have: “Can independently design, build, and maintain complex React frontends,” backed by Meta employee who have reviewed real code and systems.” • Instead of “distributed systems” as a buzzword, you have: “Can design and reason about distributed architectures at meaningful scale and can work in team” validated through a practical design exercise and a deep technical conversation with someone who already operates at that level.

Critically, those peers are not generic HR screeners. They are credible practitioners: senior and staff engineers, architects, tech leads—the kind of people you already trust when you ask internally, “Would you hire this person onto your team?”

The process has two non‑negotiable parts:

1. Real work, not artificial puzzles. Candidates demonstrate skills through practical projects, codebases, or system designs that resemble actual work, not abstract quizzes. This surfaces how they reason, structure, communicate, and trade off—things that do not show up in a bullet point, because in the real world you cannot have it all.

2. Live, technical dialogue with peers.

A structured conversation with a seasoned practitioner exposes depth: can they defend decisions, adjust under constraints, admit what they do not know, and collaborate through uncertainty? This is what day‑to‑day engineering actually feels like. They will be talking with someone who know what a team player is, someone who can tell how really capable this person is.

After the assessment is over the silent part is also evaluated and that is the personality of the candidate which is the most important thing. We as humans cannot truly tell how someone is in the inside so we will never know for sure 100%, but we can have a good understand and expectation what someone will do and is capable of.

When a candidate passes that bar, the outcome should not die in someone’s inbox. It becomes a portable, digital credential: a verifiable record that “these specific skills were evaluated at this level by these kinds of peers, using this kind of evidence.”

The personality of the candidate would highlight some of the natural features of the candidate, which can improve greatly in deciding if the person is going to function in the company’s culture. For example, one of the personality traits for someone can be that he is the best when working alone and if that is someone you want to hire, then its going to be a no-brainer for your company. Additionally, if someone is the perfect candidate technically, but lacks leadership qualities and you are looking for a team lead, that information alone can save you time and money.

For companies, this creates a skills trust layer that sits above the noise:

• Instead of starting from a pile of embellished CVs, you start from a smaller set of people whose key skills have already been pressure‑tested by experts, from all over the world. • Instead of re‑interviewing people from scratch on the basics, you focus your time on culture, role fit, domain context, and higher‑order judgment. • Instead of guessing who can really perform, you lean on structured, cumulative proof from peers whose standards you respect.

For candidates, it changes the game in the other direction:

• People from underrepresented geographies or non‑traditional paths can compete on what they can actually do, not on how close they are to your existing network. • Once a skill is verified, they are not forced to endlessly prove the same thing in slightly different formats for every new company. Their credibility starts to compound, so it can count as interview once and count as interviewed with other companies as well.

This is what a sane hiring future looks like: self‑reported claims are just the starting hypothesis; peer‑verified skills are the evidence you bet on. Instead of adding more filters on top of untrusted data, you change the data itself—from stories into proof.