If you graduated with a CS degree in the last two years and your applications are vanishing, the story you have been told (AI ate the junior seat) is mostly wrong. On June 1, 2026, New York Fed economists Natalia Emanuel, Emma Harrington and Amanda Pallais published a Liberty Street Economics analysis attributing roughly 64% of the rise in recent-graduate unemployment to remote work, not AI. That single finding changes the resume playbook, and most new grads have not caught up.
What the NY Fed actually found
Remote work explains about 64% of the recent surge in young-college-graduate unemployment, which climbed from 3.6% in March 2019 to 5.6% in March 2026. The mechanism is training: employers do not want to teach juniors over Zoom.
Emanuel, Harrington and Pallais (the same team behind the "Power of Proximity to Coworkers" NBER paper) put it plainly: "Employers may not want to hire fresh graduates onto distributed teams because it is more difficult to teach them the requisite skills from afar." The data backs it up. In "remotable" occupations, unemployment for young college grads rose about 1 percentage point from 2017-2019 to 2022-2024. For workers 29 and over in those same fields, it declined slightly. Remote work grew roughly fourfold in the same window.
The uncomfortable read: seniors already have the tacit knowledge remote settings fail to transmit. Juniors need proximity to build it. When a company goes distributed, the junior is the first hire it stops making.
From Emanuel, Harrington and Pallais at Liberty Street Economics, June 1, 2026.
Why CS majors polling worse than philosophy is a composition artifact
CS majors sit at 6.1% unemployment and computer engineering at 7.5% on the NY Fed's dashboard, higher than philosophy, but that gap is not about degree quality. It is about occupation choice, and it is fixable at the resume layer.
CS grads overwhelmingly apply into remotable occupations: generic SWE, front-end, backend microservices, anything that ships over a pull request. Philosophy grads scatter into in-person work by default: teaching, operations, sales, admin. Under the NY Fed mechanism, the philosophy grad ends up in a job employers are still willing to train juniors for. The CS grad ends up in the exact lane employers most avoid staffing with juniors.
The takeaway is not "study philosophy." It is that occupation targeting drives the outcome, not the diploma. A CS degree aimed at the right occupation is still the best hand on the table. Aimed at a remote generalist SWE listing in 2026, it is a losing ticket.
The junior seat did not get automated. It got given to a laid-off senior.
New-grad hiring at Big Tech is down more than 50% from 2019 and new grads are just 7% of Big Tech hires, per SignalFire's 2026 State of Talent report. But the seats that remain are increasingly being taken by laid-off staff engineers accepting a pay cut for a year of stability.
Total engineering hiring across the tech majors (Alphabet, Meta, Apple, Amazon, Microsoft, Netflix, Nvidia, Tesla, Uber, Airbnb, Block, Stripe) is only down 11% versus a 25% overall headcount drop. Engineers went from 46% of new hires in 2019 to 55% in 2025. Companies are still hiring engineers. They are just hiring senior ones for cheap.
That means your competition on a remote junior SWE req is not another new grad. It is an ex-Meta staff engineer with 12 years of shipped systems willing to take L4 comp. You cannot beat that person on price, credentials, or velocity. You have to compete on something they refuse to put on their resume.
Your differentiator cannot be "I do junior work cheaper." It has to be something a former Meta staff engineer would not sign up for.
The forward-deployed arbitrage
The counter-move is repositioning toward proximity-heavy roles that are growing, not shrinking: forward-deployed engineer (FDE), solutions engineer, sales engineer, on-site startup gigs. These are apprenticeship-visible, in-person, and structurally shielded from the training-from-afar problem the NY Fed identified.
SignalFire's role-level data tells the story:
- Forward-deployed engineers: up more than 30% since ChatGPT launched in 2022
- Sales engineers: up 11%
- AI/ML engineers: up 39%
- Research engineers: up 28%
- Front-end engineers: down 25%
Every one of the growing roles requires either deep specialization (AI/ML, research) or customer proximity (FDE, SE). The shrinking role is the one most easily done from anywhere with a laptop. That is not a coincidence.
Here is the exclusive number that operationalizes the reposition. In Refolk's index of professional profiles, the generic Software Engineer pool is roughly 9.3x larger than the combined Forward Deployed / Solutions / Sales Engineer pool.
| Role / segment | Figure | Source |
|---|---|---|
| U.S. "Software Engineer" title holders | ~348,406 | Refolk's index |
| U.S. Forward Deployed / Solutions / Sales Engineers | ~37,421 | Refolk's index |
| Competition density: generalist vs proximity lane | ~9.3x higher in generalist | Derived |
| Big Tech new-grad hiring vs 2019 | −50% to −65% | SignalFire 2026 |
| Engineering share of Big Tech hires 2019 → 2025 | 46% → 55% | SignalFire 2026 |
| Front-end roles since 2022 | −25% | SignalFire 2026 |
| Forward-deployed roles since 2022 | +30% | SignalFire 2026 |
The generic remote SWE lane is where every unemployed new grad is fighting an ex-staff-engineer. The FDE lane is roughly an order of magnitude thinner and growing 30%+. That is the arbitrage.
Derived from Refolk's index of U.S. professional profiles.
Where to actually apply
Target companies actively hiring forward-deployed and solutions engineers, most of them in-person or hybrid in SF, NYC, or Austin. Refolk's index surfaces these as the top hirers in the segment:
- Northslope Technologies
- Cyberhaven
- Gecko Robotics
- Modal
- Amp Code
- Island
- Roboflow
Notice what these companies share: mid-stage, technical product, enterprise or industrial customers, and a culture that expects engineers to sit next to customers or in a physical office. That is exactly the "training-visible" environment the NY Fed data says still hires juniors. Geographic concentration matters too. San Francisco is the top market for FDE roles, New York is second, Austin third. If your resume header says "Open to Remote," you are opting out of the growing lane.
How to rewrite the resume
The reposition is a resume problem more than a skills problem. A CS grad with a summer internship and a couple of side projects already has the raw material for an FDE-track resume. It just needs to be aimed correctly.
Do this:
- Rewrite the summary line for proximity. Not "Full-stack engineer open to remote roles." Try "Systems generalist based in SF, comfortable with customer travel and on-site deployments."
- Lead every bullet with an ambiguous-scope, customer-adjacent story. The hackathon where you shipped a demo in front of the sponsor engineer. The internship where you sat with the account team. The class project where you interviewed real users.
- Kill the front-end framing. If your last three bullets are React, Tailwind, and Next.js, you are volunteering into the −25% bucket. Reframe the same work as "integration," "prototyping in front of stakeholders," or "customer-facing tooling."
- Name a niche stack. Modal, Roboflow, Cyberhaven, Gecko: each has a stack an ex-staff-engineer will not list. Kubernetes on the edge, computer vision pipelines, data lineage, robotics middleware. Pick one and go deep.
- Drop generic "AI tools" bullets. Every grad says they use Cursor. That is not a differentiator. What is: shipped work Copilot cannot do alone, meaning customer-facing, ambiguous-scope, systems-integration projects.
The tailoring problem is where most new grads stall. You need a differently framed resume for every FDE or SE posting, because the vocabulary shifts by company. That is the exact work Refolk takes off you: paste the posting, get your own history back rewritten to match its language, with the cover letter drafted and a fit score telling you whether it is even worth applying.
The AI story, corrected
AI is a real force in the entry-level market, but at the task level, not the job level. That distinction matters for your resume.
New role starts for people with less than one year of post-graduate experience fell 50% between 2019 and 2024. That is where AI pressure is clearest: glue code, boilerplate, simple debugging, narrow tickets. Those tasks used to be the paid on-ramp for juniors. Tools can now absorb a lot of that work, so the industry is becoming less willing to pay for training.
ServiceNow CEO Bill McDermott has predicted new-grad unemployment could hit the mid-30s within a couple of years because agents will do the work. Even if you think that is bluster (the NY Fed data suggests it is), the underlying observation is right: the boilerplate on-ramp is closing.
The response is not to fight AI on the resume. It is to aim at work AI cannot do alone. Customer discovery. On-site debugging of a customer's actual data pipeline. Turning a vague executive request into a shipped integration in two weeks. Those are FDE and SE outputs. They are also work an ex-Meta staff engineer typically refuses to do.
The reposition in one sentence
If you are a 2024, 2025, or 2026 CS grad with a stalled search, stop applying to remote generalist SWE jobs and start applying to in-person FDE, SE, and mid-stage startup roles in SF, NYC, or Austin.
That is the entire play. The NY Fed says 64% of your problem is the lane you are in. SignalFire says the lane you should be in is growing 30%+. Refolk's index says it is 9x less crowded. The resume rewrite follows from that: proximity, ambiguity, customer-adjacent, niche stack. If the header of your resume still says "Remote preferred," you are optimizing for the exact roles employers have decided juniors cannot be taught to do.
FAQ
Is the computer science unemployment rate really higher than philosophy?
Yes, on the NY Fed's labor market dashboard for recent college graduates, CS majors sit at 6.1% and computer engineering at 7.5%, both above philosophy. But the gap is a composition artifact. CS grads overwhelmingly target remotable occupations, which is the exact lane the NY Fed identifies as where junior hiring has frozen. Philosophy grads default into in-person work like teaching, operations, and sales, where employers still train juniors. Change the target occupation and the degree-level gap largely disappears.
What is a forward-deployed engineer and why should a new grad apply for it?
A forward-deployed engineer (FDE) is a technical hire who sits with customers, often on-site, and adapts a company's product to that customer's actual data, systems, and workflows. It is closer to solutions engineering than pure SWE, and it is booming: SignalFire shows the role up more than 30% since 2022. For new grads, the appeal is threefold. It is in-person (so employers will train you), the pool is roughly 9x smaller than generic SWE, and ex-staff-engineers usually refuse the customer travel involved.
Should I put "AI tools" or "Cursor" on my new grad resume in 2026?
Not as a differentiator. Every new grad now lists Copilot, Cursor, or Claude, so it does not signal anything. What actually helps is a specific example of work you shipped that required human judgment on top of AI output: a customer requirement you translated into a spec, an ambiguous prompt you decomposed, or an integration where you resolved data model mismatches an autocomplete tool could not. Frame the work, not the tool.
How do I tailor my resume to FDE and solutions engineering postings without spending hours on each one?
The FDE and SE vocabulary shifts a lot by company (Modal, Cyberhaven, and Gecko Robotics each describe the same core job differently), so a single generic resume underperforms badly on this lane. Refolk reads your history once, then rewrites your resume for every posting in that posting's language, drafts the cover letter, and scores how well you actually fit. That way you can put your hour into researching the company, not reformatting bullets.