Almost hired a developer pretending to be someone they’re not. Be careful

That’s a useful reminder that a polished CV and rehearsed answers are not enough. I’d also verify identity, ask for a live walkthrough of a real past project, and give a small paid task based on the actual codebase before granting any production access. The biggest red flag is usually not the accent or camera issue, but the inability to explain specific technical decisions in depth.

This is becoming a serious problem in remote hiring, especially when technical interviews, portfolios, identities, and even live conversations can be manipulated. The most valuable part of this story would be the exact signal that exposed the candidate and which verification step would have caught it earlier. Many founders know they should verify people, but they do not know what a practical process looks like without making every legitimate candidate feel distrusted. Did this experience lead to a repeatable hiring checklist or become part of the IbexAI product itself?

The LinkedIn profile being only 6 months old is the detail that would've worried me most, tbh — that's a much harder thing to fake convincingly than an accent or a CV. Did you end up checking if his portfolio/past "companies" were even real, or was the interview enough to kill it for you?

This looks really promising, Matt. LinkedIn lead gen is one of the highest-ROI channels when done right, but it's brutally time-consuming to do manually. An AI agent that surfaces high-intent prospects (people actively engaging with relevant content) instead of just spraying job titles is exactly what a lot of solo founders and small teams need.

A few quick questions from someone who's done a lot of LinkedIn outreach:

• How do you define/score "high-intent" signals?

• What's the current delivery format — daily list of profiles + context, or does it go further (draft messages, etc.)?

• Any early results or case studies from users so far?

Congrats on building this out of the Harvard Innovation Lab. Would love to see how it performs for bootstrapped founders specifically.

Wishing you strong traction!

It’s becoming a norm now due to AI advancements

People are trying to cheat a lot

That's a good reminder that technical skills alone aren't enough. A short live coding task and a few questions about past projects can reveal a lot.

That is a massive bullet dodged. Identity fraud and "bait-and-switch" tactics in freelance hiring have become incredibly common, and your instinct to trust your gut was spot on.

A few key takeaways from your experience:

• The Live Video Filter: Always make video a requirement. Resisting the camera or having sudden "technical difficulties" is the easiest red flag to spot early on.

• Ditch the Script: Standard interview questions can be easily memorized or read off a screen. Asking situational, company-specific technical problems forces candidates to think on their feet and proves their actual skill level.

• Cross-Reference Timelines: A LinkedIn profile that is only a few months old but claims 5+ years of senior experience is an immediate warning sign.

Hopping on a quick live call and asking unscripted questions saved you a ton of wasted cash and potential security risks to your SaaS codebase.

Yeah, the “reading from a script” detail is the real tell, more than the accent or camera. I deal with vetting people constantly on the admin/personnel side of my job, and the pattern is always the same: someone can talk fluently about architecture or big-picture decisions in the abstract, but they fall apart on the boring day-to-day minutiae real experience burns into you — what ticket system did you use, what did your actual standups look like, who did you escalate to when something broke at 2am. Nobody fakes that level of mundane detail convincingly. Might be a useful screening question alongside the live coding test everyone’s suggesting: ask about the boring parts of the job, not the impressive parts.

That's wild

"It's wild that his LinkedIn profile was only created 6 months ago—that's usually the final piece of the puzzle that confirms a burner identity or agency-managed proxy setup. When people fabricate 5+ years of intense experience, they rarely have the digital paper trail or older, organic professional network connections to back it up on social platforms.

Kudos to you for sharing this with the community; it's a huge heads-up for other solo founders who are short on time and trying to hire quickly. Did you end up reporting the profile on the freelance platform, or are you just sticking to strict, camera-on live coding tests to instantly filter these out moving forward?

One thing this highlights is how much hiring systems still depend on surface-level trust signals. A polished CV, a convincing LinkedIn profile, a familiar company name, and a technically competent interview can create the impression that verification has already happened when it often has not.

The risk is not limited to hiring the wrong person. A fake candidate may gain access to internal conversations, private repositories, infrastructure, customer data, or business strategy before anyone notices the inconsistencies. That makes identity verification part of security, not just recruitment administration.

At the same time, companies have to avoid turning every remote interview into an interrogation. The process should be strong but professional. Consistent identity checks, verified references, live collaboration, limited initial permissions, and documented onboarding can reduce risk without punishing legitimate candidates.

I would be especially interested in the sequence of events. Did the candidate make one major mistake, or did several small inconsistencies gradually create suspicion?

This is a genuinely important warning, especially now that remote hiring, AI-generated profiles, deepfake video calls, and outsourced interviews are becoming much easier to coordinate.

The difficult part is that the candidate may still appear technically strong. They can have a polished GitHub profile, answer common interview questions correctly, and even complete a take-home assignment if another person is helping behind the scenes. By the time the inconsistencies become obvious, the company may already have shared internal documentation, customer information, credentials, or access to production systems.

A few practical checks seem more important than ever: asking the candidate to explain a recent project in detail, reviewing code together live, changing requirements during the technical session, and verifying references independently.

I would also avoid giving broad access immediately after hiring. A gradual access model during the first few weeks can protect the company without treating every new hire with suspicion.

What was the first detail that made you realize the candidate might not be who they claimed to be?

That’s a scary situation. What was the first sign that made you question their identity? A practical verification checklist would be really useful for other founders.

Great insight. I'm building Nexus Automation Agency focusing on data-driven intelligence. I’ve noticed similar challenges in finding high-intent partners rather than just freelancers. How do you guys filter for genuine technical partners?

This is a good reminder that a polished LinkedIn and a clean CV don't substitute for testing whether someone can reason through a real problem live. The actual tell in your story wasn't the accent or the camera — it was that he could describe past roles fluently but fell apart on anything specific to your company. That's usually the fastest way to catch it: ask what they'd do given your actual constraints, not what's on their resume.

Stories like this remind me that hiring is about more than finding talent. It is also about finding someone you can trust. Taking extra time at the start is much better than fixing bigger problems later.

I hired engineers on six continents for two decades and the fake-CV problem predates AI, it has just gotten cheaper to run. The screen that never failed me: ask what broke in their last project and what they did about it, because invented experience has successes but no scars. This is also often organized fraud rather than one ambitious freelancer, so verify identity through the payment rails too, the payee name has to match the person you interviewed.

This is becoming way too common lately. What works best for me is skipping the traditional resume roast and jumping straight into a short live code walkthrough. Ask them to explain a specific design choice in their public GitHub repo. Fake devs usually freeze up instantly when they cannot read off a secondary screen.

The "Soham Parekh moment" is real 😅. Resumes are getting easier to fake, but engineering depth is still hard to fake. That's why companies like GeekyAnts emphasize practical technical interviews and problem-solving over polished CVs.

You did good

The scripted answers on past roles plus LinkedIn account age are useful catches, but curious what would have happened if he'd been better prepared. The underlying fraud vector, using someone else's CV and project history, is hard to detect if the person has done enough research on the identity they're using. Is there a verification step that would actually catch that beyond interview performance?

Filipino dev here, and this stings to read — that "good experiences with Filipino freelancers" reputation took our community years to build, and impersonators are free-riding on it. For what it's worth, a vetting step that works better than accents or cameras: a short paid trial task on your actual codebase. Scripted candidates fall apart the moment the work is real and specific, and legit devs are usually happy to do it because it beats another interview round.

This is exactly the hidden cost of hiring nobody talks about —you're not just paying for the work, you're paying with your time to vet, verify, and catch red flags like this. It's why for small recurring tasks I lean toward a flat-rate service model instead of individual freelancer hires one point of accountability instead of re-vetting a new person every time. Glad you caught it before it cost you more than an interview slot."

That’s a scary situation — almost hiring someone pretending to be a developer. It shows how important verification and trust are in remote hiring. Do you think technical tests or trial projects are the best way to catch this early?

It's surprising how convincing a profile can look. A short live coding session or discussing past projects in depth usually says a lot.

Good reminder. A strong CV isn't enough always verify identity and assess real problem-solving skills, not just rehearsed answers.

The angle nobody mentioned yet: reviewing code is harder to fake than writing it. Hand the candidate a small pull request with one subtle planted bug and ask them to talk through it live. Anyone can produce code all day with AI now, but reasoning out loud about someone else's code exposes real judgment in minutes. And seconding the payment rails point, a lot of these are organized operations rather than one ambitious freelancer, so the payee name check catches more than the interview does.

The specific-company questions were useful; accent and appearance are not reliable identity checks. I would make the process repeatable: verify work history through a company-controlled contact, run a short paid task with live reasoning, and give the contractor least-privilege access until that task is reviewed. That catches scripted experience without turning nationality guesses into a hiring signal.