Why I Built Jobsolv: The Job Search Problem Nobody Was Solving

Why I Built Jobsolv: The Job Search Problem Nobody Was Solving

In March 2023, Silicon Valley Bank collapsed. I had worked there. I watched thousands of talented people -- my former colleagues, people in my network -- suddenly thrown into a job market that was fundamentally broken.

Not broken in the way most people think. The problem wasn't that there were no jobs. The problem was that the system connecting people to jobs was designed badly, and nobody seemed interested in fixing it.

That realization is why Jobsolv exists.

The Problem: A System Built for Machines, Not People

If you've searched for a job recently, you already know the pain. But you might not know why it's so painful.

Keyword Matching Doesn't Work

Most applicant tracking systems (ATS) filter resumes using keyword matching. Your resume has the right keywords? You move forward. It doesn't? You get rejected. Doesn't matter if you're the most qualified person who applied.

I've seen this from both sides. As a hiring manager building Jobsolv's small, effective startup team, I watched the ATS filter out candidates who would have been excellent fits -- simply because they described their experience differently than the job description expected. A candidate who wrote "A/B testing" when the JD said "experimentation" would get filtered. Same skill. Different words. Rejected.

The system is optimized for efficiency, not accuracy. It's great at processing thousands of applications quickly. It's terrible at identifying the best candidates.

Applications Are a Black Hole

The average job seeker spends hours customizing each application. They tailor their resume, write a cover letter, fill in redundant forms that ask for the same information already on their resume. Then they submit -- and hear nothing.

The data is depressing: the average online job application has roughly a 2-3% chance of leading to an interview. That means for every 100 hours a job seeker spends applying, they might get 2-3 conversations. Most of those won't lead to offers.

This isn't just inefficient. It's psychologically devastating. Job search is already one of the most stressful experiences a person can go through. When the process itself is designed to waste your time and give you no feedback, it compounds the psychological toll in ways that most people underestimate.

Existing Solutions Missed the Point

When I started researching the space, the existing solutions fell into two categories:

Job boards that aggregated listings but did nothing to help you actually get the job. They made the top of the funnel bigger without improving the conversion rate.

Resume builders that helped you create a nice-looking document but didn't address the fundamental problem: your resume needs to match what each specific employer is looking for, and that changes with every application.

Nobody was solving the actual problem: the disconnect between a qualified candidate and the automated system deciding their fate.

The Behavioral Design Approach

My background in behavioral economics and experimentation gave me a different lens on the problem.

Most people think of job search as a logistics problem: find listings, send applications, wait. But it's actually a behavioral problem -- on both sides.

On the employer side: hiring managers have cognitive biases. They anchor on specific keywords, they're influenced by resume formatting, they make snap judgments based on pattern matching. The ATS encodes these biases into algorithms.

On the job seeker side: people make systematically poor decisions under the stress of unemployment. They apply too broadly (spray and pray), they undersell themselves, they spend time on low-probability applications while ignoring better-fit opportunities.

I realized that the solution needed to address the behavioral dynamics, not just the technical mechanics. That meant:

1. Understanding what the ATS is actually looking for -- not just keywords, but structure, formatting, and context.
2. Automating the tailoring process so each application is optimized without requiring hours of manual work.
3. Matching people to jobs they'd actually accept, not just jobs that match keywords on their resume.
4. Reducing the psychological burden by making the process faster, more transparent, and more successful.

Validating With Services First

Before building any technology, I needed to validate the core hypothesis: could we dramatically improve job search outcomes by applying behavioral design and systematic optimization to the application process?

We started with a concierge service. Real humans, working with real job seekers, manually applying the methodology.

The results were clear. We worked with 26 initial clients, charging $2,000 to $3,000 each. Our interview success rate hit 92.3%. That's our Signature Service -- not a marketing claim, but a validated outcome from doing the work manually before automating it.

Those numbers told me two things: the approach worked, and there was a market willing to pay meaningful money for it. The next question was whether we could scale it.

Building the AI

The technology behind Jobsolv was designed to replicate what worked in our service model, at scale.

Automated Resume Tailoring

The core product takes a job seeker's base resume and the target job description, then generates a tailored version that's optimized for both ATS systems and human reviewers. It's not just swapping keywords -- it restructures accomplishments, adjusts emphasis, and reformats to match what each specific role requires.

This is where the behavioral economics background mattered most. We're not just matching keywords. We're applying principles of framing, anchoring, and information architecture to present each candidate in the strongest possible light for each specific role.

Job-Matching Workflows

We built workflows that go beyond simple keyword matching. Instead of showing you every "marketing analyst" job on the internet, we identify roles where your actual experience, skills, and career trajectory are genuine fits -- roles you'd actually accept if offered.

This distinction matters more than most people realize. A huge source of job search frustration is applying to roles that look right on paper but are wrong in practice -- wrong level, wrong culture, wrong growth trajectory. Our matching accounts for dimensions that simple keyword search can't capture.

ATS-Compliant Formatting

This sounds mundane, but it's critical. Many ATS systems can't parse creative resume formats -- columns, tables, graphics, headers in unusual positions. We ensure every resume we generate is structured in a way that ATS systems can read correctly while still looking professional to human reviewers.

You'd be surprised how many qualified candidates are rejected because their beautifully designed resume is unreadable to the software screening it.

Where Jobsolv Is Now

From those first 26 service clients, Jobsolv has grown to a platform serving 35,000+ job seekers. We've generated $80,000+ in revenue with zero advertising spend -- entirely through word of mouth, content marketing, and the product delivering genuine results.

At peak, we ran a small, effective cross-functional team spanning engineering, data science, content, operations, and customer success. We acquired two remote job boards to expand our reach. We built partnerships across the career services ecosystem.

The numbers I'm proudest of aren't the revenue or user count. They're the outcomes: the 92.3% interview success rate from our Signature Service, and the thousands of people who've used our platform to land roles they're genuinely excited about.

The Lessons From Building in Public

Building Jobsolv taught me things that my entire career in analytics and experimentation hadn't.

The Unsexy Problems Are the Valuable Ones

ATS parsing, resume formatting, keyword optimization -- none of this is glamorous. Nobody writes breathless blog posts about the future of resume formatting. But these unsexy, infrastructure-level problems are where the actual value lives.

The most impactful thing we do isn't impressive from a technology standpoint. It's making sure a qualified candidate's resume gets past the first automated screen. That's it. And it changes people's lives.

Services Before Software

Starting with manual services before building technology was the best decision we made. It validated the approach, generated revenue, taught us what customers actually needed (as opposed to what we assumed they needed), and gave us training data for the AI.

If you're building in the job search or career tech space, I'd strongly recommend this path. The temptation is to build the product first. Resist it. Do it manually. Learn what works. Then automate.

Job Search Is Deeply Emotional

This was the biggest lesson. I came into this as an analytics and experimentation person. I thought the problem was technical -- fix the keyword matching, optimize the resume, improve the application process.

The problem is emotional. People searching for jobs are stressed, often scared, frequently dealing with identity questions ("Who am I if I'm not employed?"). The product needs to address that emotional reality, not just the technical workflow.

We redesigned several features after realizing that some of our "efficient" workflows were actually increasing user anxiety. Showing people a list of 50 jobs they should apply to isn't helpful -- it's overwhelming. Showing them 5 genuinely strong matches with clear next steps? That's useful.

The Vision

Jobsolv started as a response to a crisis -- the SVB collapse and the broken job market it exposed. But the problem we're solving isn't crisis-specific. It's structural.

The vision is straightforward: AI that actually matches people to roles they'd accept, not just roles that share keywords with their resume. A job search process that respects people's time, reduces psychological burden, and produces genuine outcomes.

We're not trying to replace human judgment in hiring. We're trying to ensure that human judgment actually gets applied -- that qualified candidates make it past the automated filters to the point where a human being actually looks at their application.

That's the job search problem nobody was solving. And it's why I built Jobsolv.

If you want to learn more about my background and the thinking behind the product, visit atticusli.com. If you're currently searching for a job and want to see what Jobsolv can do, check out jobsolv.com.

Atticus Li is the founder and CEO of Jobsolv. An experienced hiring manager for both startups and enterprise — he's sat on hiring teams at Fortune 500 companies and venture-backed startups, directly vetting resumes, running interviews, and helping hire 150+ candidates. He mentors ambitious recent grads, early-career professionals, and nontraditional candidates breaking into data, analytics, marketing analytics, business intelligence, and data science. He holds certifications in behavioral economics from Mindworx (Ogilvy Group UK) and CRO from CXL Institute. Reach out at [email protected].