Closed beta. Two students, a handful of friends, and a waitlist we did not plan on.
Stop applying. Start campaigning.Qualified is not the same as visible.

Job search is rigged.
We're here to fix it.

The college job search isn't hard because jobs disappeared. It's hard because the machinery between a qualified student and a hiring manager is optimized to discard people at scale. The system isn't broken by accident. It's broken by design.

Searchgrunt is the job search operating system for students, who refuse to be rejected by broken systems, built by students, who won’t allow a line of if/then code, written a decade ago, to decide their future.

Read the manifesto ↓
One search / many angles
searchgrunt.app / today

Today

The morning brief. Three ranked moves, the reason each one is ranked, and a timer that ends the session before the session ends you.

/public/screens/hero-today.png
your strategy,
not theirs ↗
No more
spray
& pray
01 / Manifesto
An unfair system
is still a system.
We bypass it.

You’re competing
with 108 other people.
Per opening.

The story is that students lose because they are unready, ungrateful, or insufficiently “passionate.”The real story nobody talks about is structural.

Companies flood the market with postings, sometimes for jobs that do not exist, then hide the hiring manager behind an applicant tracking system and a recruiter who in most cases has never worked in the function being hired.

The number everyone repeats is that 75% of résumés never reach a human. Nobody has ever produced the study behind it, so treat it as folklore. A study that does exist comes from Harvard Business School and Accenture who found that 88% of employers acknowledge their own screening process pushes qualified candidates out. Not because those people cannot do the work. Because the software, and the humans feeding it, consistently and repeatedly get it wrong. Hidden Workers ↗

Recruiters then make first-cut decisions they are structurally unequipped to make. Research found that when technical recruiters rejected candidates they were convinced would never pass a technical interview, those same candidates passed roughly half the time anyway. The people deciding who is qualified to interview are, statistically, about as accurate as a coin flip. The coin-flip study ↗

Universities offer career centers that are, with some honorable exceptions, underpowered and oddly staffed. They are often filled with well-meaning advisors who have never sat in the rooms students are trying to enter. Handshake logins. Résumé templates. A spring fair with tote bags. Advice that has not changed since the late 1990s, printed on a poster. A decade ago, Gallup found that only 17% of graduates called the career center “helpful” in finding a job. Today, that number is much lower. Gallup, 2016 ↗

Some data would help. None of it reaches the student. No honest view of which sources produce interviews. No map of the alumnus two years ahead of you in the target group. No honest score of stretch versus fit. No record of whether your neighbor who works at the company is a warmer path than another keyword-optimized PDF.

So students do the only thing
the system taught them:apply more, feel worse,
and call it a search.

The keyword tax

Your work, flattened into nouns.

The filter rewards exact phrasing, familiar employers and lucky timing. It has no opinion about whether you can do the job, because it was never built to evaluate your abilities. It was built to evaluate how well you stuff your résumé with keywords.

The recruiter lottery

Fast judgment, borrowed context.

A generalist (often not much older than the student) screens for a function they have never worked in, at volume, on a clock. The people closest to the work keep missing qualified candidates the top of the hiring funnel already threw away.

The campus ritual

A workshop is not a strategy.

Some career offices are excellent. Most, sadly, are not. Under-funded, under-resourced, under-appreciated and overwhelmed. At most schools, the ratio of students-per-advisor is well over 1000-1. That’s why the résumé workshops and career fairs are identical to the ones your parents and even grandparents attended.

The data vacuum

Plenty of activity. No signal.

You count applications because applications are the only thing you can see. Which source produced an interview, which path ran warm, whether the last forty hours moved anything? None of that info ever reaches the applicant. The system was designed to be opaque.

50%

of candidates doubt the job they are applying to is real

Only half trusted that the opening on the other side of the form actually existed. In a challenging economic environment, companies are more concerned about projecting their growth and health than wasting the precious time and energy of hundreds or even thousands of applicants.

Gartner, July 2025 · ~3,000 candidates ↗
38%

were ghosted by an employer in a single year

Not a rejection. Complete and total silence. In many cases, not even an acknowledgement of receiving an application. A third of candidates now assume they have been ghosted after one week, because that has become the reasonable assumption.

Criteria, 2024 Candidate Experience Report · 2,516 candidates ↗
43%

of graduating seniors had even one offer in hand

An average of 0.79 offers per graduate. The days of deciding between offers are long gone. Interns hoping to get full-time offers after the summer shouldn’t hold their breath. Nearly half of students who had paid internships had ZERO job offers at graduation.

NACE 2026 Student Survey · 1,860 graduating seniors ↗
One more figure, for context

Employers plan to hire more students this year.

Employers were projecting a 5.6% increase in hiring this year.
Demand went up.
Applications per opening tripled anyway.
Jobs are not the scarce thing.
A path to a human being who can say yes is the scarce thing.

Everything we built
was to find the paths.

Field note 001
“Applied to 83 jobs” is a fact. It has never once been a plan.
02 / Solution

One search.
Run like a campaign.

Searchgrunt answers five questions continuously, so the answer is waiting for you instead of being reconstructed from a spreadsheet at eleven at night. Discovery runs while you sleep. Ranking happens against your criteria. Access gets mapped before you need it. Nothing is sent without you.

Engine i

Discover

Conventional and not: job boards, company portals, mailing lists, blog posts, research groups, fellowships, government pathways, hiring intent signals. Plus anything you forward in from Mom, your roommate or a GroupMe thread. We ingest it, dedupe it, enrich it.

// What exists that I would never find on my own?
Engine ii

Decide

Priority scoring on role fit, skill fit, freshness, deadline, and effort weighed against how strong your access actually is. No more doom-scrolling three hundred tabs to feel productive. No more wasting time applying to a stale posting that may already have 400 applicants.

// Which ones are worthy of my time?
Engine iii

Direct

This is the unfair advantage. Parents, alumni, neighbors, teammates, camp friends, professors, former interns, hiring managers and your mom’s middle school friend who is now a CFO. We map affinity paths you did not know you had.

// Who can get me in front of a human?
Engine iv

Differentiate

A generic cover letter written by Claude? Please. We make you a LinkedIn legend and build assets a recruiter will ignore but a hiring manager will LOVE: a teardown, a data analysis, a policy memo, tailored to real business challenges.

// What makes me impossible to ignore?
Engine v

Drive

Today tells you exactly what to do next. Automatic follow-ups, deadlines, calendar. Plus real analytics: your alumni outreach converts five times better than cold recruiter outreach, so do more of that.

// What is the single best thing to do right now?

Every serious opportunity opens two tracks at once.

The application track goes through the employer's real system, before the deadline, via the mandatory online submission.

The access track goes through people: an alum two years ahead of you on that team, a professor who used to work there, your father's boss who knows the CEO.

Both tracks are visible on the same card. When one stalls, the system says so and recommends the next play with a no-BS estimate of what it will cost you in time.

searchgrunt.app / campaigns

Campaigns

Dual-stream tracking. Application status on one axis, access status on the other, with the recommended next play and what it costs in minutes.

/public/screens/campaigns.png
Fig. 01Application track and access track, on one card.
Field note 002
“Keep checking job boards” is advice from people who have not had to find a job in twenty years.They mean well, but they are picturing a world that no longer exists.
03 / Product

What it actually does.

We built what we wished we had last year. A system that levels the playing field.

01
Today

The morning brief.

Open the app and get three moves, ranked, with the reasoning visible. Not a dashboard of everything. Three things, in order, because you have six hours a week for this and not sixty.

  • —Since your last visit. New matches, deadline shifts, replies and inbound leads, summarized rather than listed.
  • —45-minute search sprint. A timed session with a queue built for it. Mark done, next item loads.
  • —Ranked by timing, fit and access. Never by the company who paid to have their job appear first in the list.
searchgrunt.app / today

Today

Return brief, focus queue, sprint timer.

/public/screens/today.png
searchgrunt.app / opportunities

Opportunities

Discovery and triage. Role fit, skill fit and stretch, scored separately.

/public/screens/opportunities.png
Opportunities

Discovery and triage.

The harvest engine pulls continuously from employer applicant systems, program feeds and registries, normalizes the rows, and scores each one against the search lenses you defined. You triage. The machine does not decide for you.

  • —Greenhouse, Lever, Ashby, RSS and Atom feeds, federal and parliamentary program listings, curated registries and public APIs.
  • —Three scores, not one. Role fit, skill fit and stretch, each shown separately so a stretch is a choice rather than a surprise.
  • —Domestic and international streams side by side, because one of us is applying from Scotland.
  • —The canonical link. You apply through the employer's real system, never a scraper's copy of it.
02
03
Campaigns

Two ways in, tracked as one.

A campaign is a company, a role, or a whole industry, with an outcome attached. It tracks the paperwork and the people at the same time, and it tells you which of the two is currently stuck.

  • —Dual-stream status. Application status and access status move independently, because they do in real life.
  • —Named plays. Parent or alum introduction, direct recruiter inquiry, informational conversation, post-application follow-up. Each with a best-use note and an effort estimate.
  • —Campaign health. Cold campaigns surface themselves instead of quietly dying in row 47 of a spreadsheet.
  • —Timeline. Every send, reply and silence, logged, so follow-up is a decision rather than a memory test.
searchgrunt.app / campaigns

Campaigns

Board view, campaign health, recommended next action.

/public/screens/campaigns.png
searchgrunt.app / network

Network

The affinity graph. Every edge you actually have, ranked by warmth.

/public/screens/network.png
Network

The affinity graph.

Everyone says use your network. Nobody tells you which twelve of the four hundred people you vaguely know are connected to the eight companies you actually want. This does.

  • —Edges that exist. University, secondary school, hometown, clubs, fraternities and sororities, places of worship, summer camps, faculty, family, friends of family, previous employers.
  • —Mapped onto targets. Contacts attach to campaigns, so a person is always attached to a reason.
  • —Warmth ranking. A third-degree alum who runs a business unit beats a first-degree cousin three industries away.
  • —Outreach drafted in your voice. Reviewed by you, sent by you, from your own address.
04
05
Contributor inbox

Let the adults help.

Your parents want to help and have no idea how, so they forward you a link at midnight and ask about it on Sunday. Give them a private submission link instead. Professors, mentors, family friends, the neighbor who knows someone: they all get your personal submission link. Get your friends and family working on your job search.

  • —A token link, no account, no login, works from a phone in a parking lot.
  • —Entities extracted automatically. Company, role, person, deadline and source pulled out of whatever they paste.
  • —Your review queue is the gate. Nothing a contributor sends enters your search until you say so.
  • —Advisors too. Professors, mentors, family friends, the neighbour who knows someone at the agency.
searchgrunt.app / submit / …

Contributor inbox

Inbound leads from parents and advisors, parsed into structured leads awaiting your review.

/public/screens/contributor-inbox.png
searchgrunt.app / postmaster

Postmaster

Post studio. Subjects researched, drafted in your voice, with the 210-character fold marked.

/public/screens/postmaster.png
Postmaster

Become someone they have already heard of.

Applying is a push. Being known is a pull. Share content worth having an opinion about in your field, drafted automatically in your own writing voice. If executives use publicists and AI to produce their LinkedIn posts, why can't you?

  • —Three sources of subject. A URL you paste and it reads, a scan of your saved topics across news and trade feeds, or angles researched against a specific target employer.
  • —Your voice, configurable, trained on writing samples rather than a generic professional register.
  • —Live preview with the fold marked at 210 characters, plus Unicode emphasis and hashtag suggestions.
  • —Copy and hand off. Nothing posts on your behalf. Ever.
06
07
Progress

Evidence instead of activity.

Application counts measure how tired you are. This measures whether the search is actually moving, and it will tell you when it is not.

  • —Response rate, by source and by path, so you learn which channels are worth the hours.
  • —Discovery lift. Roles surfaced that you would not have found through the boards you already check.
  • —Coverage. Priority campaigns that have a viable access path, versus the ones running on paperwork alone.
  • —Calendar and work sessions, so the search occupies scheduled time rather than every anxious gap in the day.
searchgrunt.app / progress

Progress

Response rate, discovery lift, access coverage, sources scanned.

/public/screens/progress.png
Field note 003
A tracker remembers what you did.It has no opinion at all about what you should do next.
04 / Unfair Advantage

Why you cannot
just do this yourself.

You could, in theory. Keep a spreadsheet, check nine boards every morning, remember who your father golfs with, and follow up on the right day. People do try. The attempt takes about fifteen hours a week and collapses in the third week of October, which is exactly when the deadlines land.

Advantage i

Access is the only lever that moves.

Every other input is capped. Your GPA is set, your experience is what it is, and your résumé is already as good as a résumé gets. The one variable still open is whether a human being sees you before the filter does. That is the variable this system works on every single day, and it is the one nobody built a tool for, because mapping a nineteen-year-old's real network isn’t a business idea VCs were rushing to fund.

Advantage ii

It compounds. Your spreadsheet does not.

Every lens you tune, every contact you confirm, every response you log makes the next week sharper. A spreadsheet is the same spreadsheet in March as it was in September, except longer and more depressing. By the time a manual search knows what is working, the season is over.

Advantage iii

Six hours a week, spent correctly.

You are a full-time student. The search competes with a degree, a job, and a life. Doing this by hand means the hours go to administration: checking boards, updating rows, remembering. Here the administration is finished before you open the laptop, and the six hours go to the three things that actually change an outcome.

What the tool does for youYou and a spreadsheetJob boardsAI résumé toolssearchgrunt
Collects roles in one place●●—Yes
Finds roles you never thought to search for———Yes
Ranks by your criteria, not the seller's●——Yes
Names a person who can open the door———Yes
Turns a parent's tip into a tracked lead———Yes
Tells you the next move, in order———Yes
Notices when a campaign goes cold———Yes
Measures whether any of it is working———Yes
Sends nothing without you reading it first●——Always
Field note 004
Hard work and a stream of accomplishments deserve more than an instant no.Especially from a bot that did not like the font on your résumé.
05 / Our Story

Two brothers.
One frustrating summer.

One was getting filtered out of engineering internships in the US. The other was getting politely ignored for strategy and policy roles across the ocean. Same machine, different accents. They compared notes and realized the battle was asymmetrical, so they built a weapon.

JR
Co-founder / systems

Jacob Rybchinin ↗

Sophomore at the University of Maryland, studying Artificial Intelligence and Information Science. Spent the year watching a stream of applications go into employer systems and come back as silence, and worked out that his résumé was not the problem. Built the harvest layer, the scoring, and most of the parts that run while everyone is asleep.

School
University of Maryland
Year
Sophomore
Owns
Ingestion, retrieval, data model
ZR
Co-founder / strategy

Zachary Rybchinin ↗

Fourth year at the University of St Andrews, a triple major working toward an MA in International Relations, Russian and Mandarin. An American navigating the British and European hiring maze, which is more complex, more bureaucratic and more heavily guarded. Developed the strategy, the campaign playbooks, and every word on this page.

School
University of St Andrews
Year
Fourth year
Owns
Access model, plays, voice

Spring was not a good season for us.

One of us had sent well over a hundred applications into engineering internship portals and heard back from a number you could count on one hand. Most did not even respond with an automated thank you for applying.

The other was applying to US and European roles from Scotland. In Europe, a US home address could not even be entered into many portals. In the US, an undergraduate claiming he was working toward an MA in three different subjects fought the ATS and baffled recruiters.

Two different countries, two different fields, two versions of the same story.

We compared our spreadsheets, ran a postmortem on the classmates who did manage to secure internships, and found some very clear patterns.

We didn’t run any sophisticated statistical analysis. Anecdotally the pattern was obvious: the vast majority of classmates who got jobs did not get them through the career center or through online applications. We hesitate to land on a number, but it is clearly well over 75% and probably north of 90%. The widely repeated claim is that seventy to eighty percent of jobs are never publicly posted. Gallup's finding that only 17% of graduates called the career center helpful points in the same direction. We think our anecdotal sample is pretty accurate.

Both of our schools have great career centers. The people there are genuinely lovely and really do want to help, though it is clear they do not fully appreciate just how powerless they are in this environment.

So we turned our anger into motivation.
Instead of lounging by the pool and rewriting our résumés for the twentieth time, we would fight back. If the process was going to treat us like rows in a table, we were going to build something that treated the process like a system with holes that can be exploited.

Zachary took the strategy. Jacob took the machinery. We argued about features and scope roughly every twenty minutes and got an MVP out anyway.

We built this for two users: us.
That constraint is why we think we actually have something. Nothing in here was built to look good in a demo or impress an investor. We built it because it is what we would use. A solution shaped around the specific problems we were having that week.

Turns out, as we already suspected, our problems were not unique. As friends and family heard what we were doing, everyone got curious. At first we just showed them screenshots. Eventually we built a small, simple login so people could create their own profile. Hiring season has not even kicked into high gear and the response has been overwhelming.

Building an app a handful of friends and family can use is straightforward. Building a stable, secure, responsive, bug-free application that anyone can use is a massive undertaking.

So we have not decided. We are, in the end, two students with degrees to finish. This is a fun app but it’s not the next Google, and whatever the media implies, it’s very rare to make any money building apps.

For now we are keeping the beta small, collecting a waitlist, and finding out how many people are as annoyed as we are… and who thinks this is worth a few dollars.

Closed beta / for now

Get on the list before
we decide what this is.

We are still working out whether this becomes a product or stays the thing two brothers built for themselves. The waitlist is how we find out. Tell us who you are, what it would be worth, and precisely which part of this process has been wasting your life. We read all of it.

Start over at the manifesto ↑
Field note 005
Stop hoping recruiters notice you. Start making them chase you.Isn’t that what the word recruit is supposed to mean?