Research Report: Tech Recruiter Exposes ATS & Common Resume Mistakes

Video

Source: Tech Recruiter Exposes ATS & Common Resume Mistakes by James Peralta

Executive Summary

This video is a candid interview clip with Donald, a former data scientist who transitioned into technical recruiting in Tokyo. Having spent five years building AI models for insurance fraud detection in Canada before starting his own data science consulting firm in Japan, Donald brings an unusually dual perspective to recruiting: he has personally experienced the frustrations of job hunting and now sees the process from the other side of the table. He works as a headhunter at a small recruitment agency, placing candidates in data science, software engineering, and adjacent roles.

The conversation centers on what actually makes a resume stand out, how ATS (Applicant Tracking Systems) work in practice, and a live critique of a junior developer's resume. Donald's core message is that a resume is not a list of responsibilities — it is a narrative tool designed to show what you can do for a company. Recruiters, especially for high-volume junior roles, often spend as few as 10 seconds on a resume before making a call. Structure, clarity, and relevance to the specific role are therefore not polish — they are prerequisites.

On the ATS front, Donald offers a useful correction to common candidate anxiety: modern ATS tools have largely moved away from rigid keyword matching and now use AI (Claude, OpenAI) to evaluate resume-to-job-description fit holistically. The real filtering danger for most candidates is platform-level pre-screening — LinkedIn's upfront questions that silently discard applications before a recruiter ever sees them.

Key Takeaways

• Tenure signals commitment: Recruiters read 2.5–3+ years at a role as evidence that a candidate can commit, deliver, and grow — not just collect experience. Job-hopping every few months is a red flag.
• Junior resumes should lead with education: If you have limited work history, put your university and GPA (if 3.6 or above) at the top. Below 3.6, leave it off entirely.
• Accomplishments beat responsibilities: Bullet points should show what you achieved, not what you were assigned. Recruiters want to see proof you can act independently, not a list of duties they'd have to manage.
• Tailor to the industry, not just the role: Keywords matter, but so does industry context. A health dashboard project is a strong signal for health-tech clients and noise for telecom clients. Match your resume's language to the company's world.
• LinkedIn's pre-screening is the real silent killer: Applications that fail LinkedIn's upfront questions (e.g., "years of experience in Python") are quietly archived — the recruiter never sees them. Lying is not the answer, but candidates should know this filter exists.
• Modern ATS uses AI, not just keywords: Platforms increasingly use LLMs to score resume-to-JD match. The implication is that coherent, well-written resumes score better than keyword-stuffed ones — write for a smart reader, not a keyword counter.
• Your resume should tell a story, bottom to top: Some recruiters start at the earliest experience and scroll up to trace a candidate's trajectory. A resume that shows a coherent arc — analytics background → dashboard work → mobile analytics — is easier to champion internally than a scattered collection of projects.

Detailed Analysis

Who Donald Is and Why His Perspective Matters

Donald's credibility in this conversation comes from having lived both sides. As a job seeker in Canada, he watched recruiters misrepresent him to clients — a frustration that eventually drove him to build his own consulting practice. When he had to hire freelance engineers for a project in Japan, he found himself surprisingly good at evaluating talent. That combination — personal experience with bad recruiting and firsthand success hiring for technical roles — motivated him to become a recruiter himself, specifically to improve the candidate experience in the Tokyo market.

He distinguishes between two recruiter archetypes: headhunters (agency-side, placing candidates across multiple clients, more diversity of opportunity) and in-house talent acquisition (embedded within a single company, focused on growing that organization). Donald is the former. This matters because headhunters are incentivized to find the best mutual fit across their client portfolio — they have more flexibility to match a candidate to the right context than an internal recruiter locked to one company's openings.

What Recruiters Actually Look For on a Resume

The highest signal for senior candidates is tenure. Donald is direct: two and a half to three years minimum per role is what serious recruiters look for. It communicates that a person can embed in a team, grow through difficulty, and contribute at depth — not just look good during a honeymoon period before bouncing.

For junior candidates, the calculus shifts. With limited work history, the resume should be structured around education first. A GPA above 3.6 belongs on the page; anything under should be omitted — it's not dishonest, it's editorial judgment about what helps you. From there, the work experience and project sections carry the weight.

The most common mistake Donald identifies across both levels: writing responsibilities instead of accomplishments. "Responsible for building a dashboard" tells a recruiter what you were assigned. "Built a health dashboard that tracked biomarker trends for clinical users" shows what you delivered. The framing shift is from "here's what I need from you" to "here's what I can do for you" — and that inversion is exactly what hiring managers want to see.

How ATS Systems Actually Work

Donald walks through the ATS landscape clearly, and it differs from what most candidates assume.

The most dangerous filter is not a software algorithm — it's LinkedIn's pre-screening questions. When you apply through LinkedIn, the platform asks upfront questions (years of experience in a language, specific certifications, etc.). If your answers don't clear the threshold, your application is silently moved to an archived folder. The recruiter has to actively click a separate tab to see it — and most don't. You may believe you submitted an application when it effectively never arrived.

Beyond LinkedIn, the major ATS platforms are Workday and Greenhouse. Both use similar scoring mechanisms. Historically, this meant keyword matching — presence of terms like "Harvard," "Stanford," or "Meta" could boost a resume's score. But Donald notes the industry has shifted: modern ATS pipelines now route resumes through LLMs (he specifically names Claude and OpenAI) to assess fit. The system asks the model how well the resume matches the job description and gets a holistic score, not a keyword count.

The practical implication for candidates: coherence and clarity matter more than keyword density. A resume written for a human reader — with clear accomplishment language, industry-relevant framing, and logical structure — will score better with an LLM evaluator than one that mechanically repeats job-posting phrases.

Live Resume Critique: What Good Looks Like

Donald reviews a resume from a junior developer attending Wilfrid Laurier University, and the critique is instructive in both directions — what works and what doesn't.

What works:

• University listed first, appropriate for the career stage
• 3.9 GPA prominently featured
• Coding languages bolded within bullet points, making them scannable at speed
• Work experience bullets focused on accomplishments (building dashboards, working with biomarkers) rather than generic task lists
• GitHub links included in the projects section — Donald calls this "validation and proof," evidence you can actually build what you're claiming

What doesn't work:

• The projects section is a wall of text: two GitHub links, seventeen coding languages stacked on the side, dense descriptions — no clear reading direction
• Project names are internal/invented ("QURE") rather than descriptive. Donald's fix: rename to something like "Airbnb Price Analytics and Geospatial Mapping" — immediately understood by anyone scanning
• The resume's story is slightly muddled: it starts in geospatial/health analytics, moves into general mobile development, and doesn't clearly signal where the candidate wants to go next

Despite the flaws, Donald rates it 8/10. The work experience and education sections are strong enough that he'd move the candidate forward for any analytics-adjacent mobile development role. The advice to get to 10: clean up the projects section's layout and sharpen the narrative about target direction.

Timestamped Topic Outline

Sources & Further Reading

• Full Stream: The full livestream this clip is sourced from is available at youtube.com/watch?v=6KzoCk1GVCw&t=3546s (linked in the video description)
• No external papers, books, or articles were referenced in this video. Donald's insights are based on his direct experience as a headhunter in Tokyo's tech recruiting market.