# Upwork Cut 37%. Its Name Shows Up on 94 Profiles per Real Employee

_Upwork's November 2026 layoffs could hit about 230 staff, but 59,156 profiles say Upwork. How to find the real marketplace, T&S and payments engineers._

- Canonical URL: https://www.refolk.ai/blog/upwork-layoffs-november-2026-find-real-employees
- Published: 2026-11-20
- Reading time: 9 min
- Keywords: Upwork layoffs November 2026, hire laid off Upwork engineers, marketplace matching engineer recruiting, trust and safety engineer sourcing, find employees of a specific company LinkedIn

If you want to hire the marketplace, trust and safety (T&S) and payments engineers Upwork is letting go, the obvious search won't find them. "Upwork" appears in the headline of 59,156 profiles in Refolk's index, about 94 for every person on its payroll, so a company filter mostly returns people who never worked there.

On November 12, 2026, Upwork disclosed that its layoff would grow from about 24% of the workforce to about 37%. That could put roughly 230 people on the market who built the search, matching, T&S and payments systems behind a marketplace with more than 18 million freelancer profiles.

## What did Upwork announce in November 2026?

Upwork's board approved a larger restructuring on November 10, 2026, and the company disclosed it two days later. The cut is now about 37% of total workforce, up from about 24% in the plan announced in May.

The key details from the filing and the coverage:

- **Scope:** about 37% of total workforce, including a "separate, independent reduction" at a subsidiary the filing does not name.
- **Cost:** pre-tax restructuring charges are now $27M to $37M, up from $16M to $23M. Upwork had booked $18.1M through September 30.
- **Timing:** mostly complete in Q1 2027 (the earlier target was Q4 2026), subject to "local-law requirements."
- **Leadership:** CFO Erica Gessert leaves on January 1.

The first wave, on May 7, 2026, cut about 145 jobs. CEO Hayden Brown's memo that day included the line "Two pizza teams are dead." The May round was about how product and engineering teams were organized. The November round reaches further.

## How many Upwork employees are actually affected?

About 233 people, if you apply 37% to the roughly 630 employees in Upwork's FY2025 10-K. The May plan at 24% worked out to about 151, so the November expansion adds roughly 82 people.

Treat those numbers as estimates. The 37% may be measured against a base that includes subsidiaries and WorkMotion, which would change the real count. Still, the order of magnitude is clear: this is a few hundred people, not thousands.

The 10-K has a second number that matters just as much: Upwork used about 1,600 independent team members on internal work during 2025. That's 2.5 contractors for every employee. Those contractors aren't counted in the 37%, but many of them did real internal work on the marketplace.

## Why does an "Upwork" search return freelancers instead of staff?

Upwork has more than 18 million freelancer profiles, and many freelancers put the platform's name in their headline or list it as their employer. That buries a staff of about 630 under tens of thousands of people who never worked for the company.

Refolk's index of professional profiles shows the scale of the problem:

| Segment | Profiles | Upwork as current employer (sample of 25) |
|---|---|---|
| Any headline mentioning "Upwork" | 59,156 | 12% |
| Freelancer-type titles + "Upwork" | 6,287 | 40% |
| Senior/manager eng, ML or T&S titles + "Upwork" | 187 | 20% |
| Actual Upwork employees (10-K, YE2025) | ~630 | n/a |
| Internal contractors in 2025 (10-K) | ~1,600 | n/a |

The first three rows come from Refolk's index. The last two come from the FY2025 10-K.

> **94 to 1** Profiles with "Upwork" in the headline for each real Upwork employee
>
> 59,156 headline mentions in Refolk's index against about 630 employees at the end of 2025.

Even if you count every contractor as well as every employee, it's still about 27 headline claimants per person who did internal work. Engineering and leadership titles make up 0.32% of all "Upwork" mentions. Freelancer-type titles alone make up 10.6%.

Narrowing to senior titles doesn't fix it either. Among the 187 profiles in Refolk's index with senior or manager-level engineering, ML or T&S titles plus "Upwork," the top locations were Bengaluru, Istanbul, Islamabad and Vietnam, not the Bay Area. The top employers included Cognizant, EY, LTIMindtree, Capgemini and IBM. Only 5 of the 25 sampled currently work at Upwork.

## Why is "current company = Upwork" the wrong filter?

Because freelancers list Upwork as their employer more often than engineers do. In Refolk's samples, 40% of freelancer-titled profiles name Upwork as their current employer, against 20% of engineering-titled ones.

The reason is how a freelancer's reputation works. Titles like "Top Rated Freelancer at Upwork" exist because the platform is where their ratings and client history live. Listing it as your employer tells clients where to find that record. It's marketing, not employment history.

So the usual way to find employees of a specific company on LinkedIn (filter by current company, then add a title) fails here in a specific way. The filter favors the people you don't want. Adding "Senior Engineer" doesn't help much either, because the senior-titled "Upwork" profiles cluster at IT services firms in Bengaluru and Istanbul.

What you need are signals freelancers rarely have and real staff almost always do.

## How do you find the real Upwork engineers?

Combine location, product names and company history. Any one of these is weak on its own. Two or three together pick out the people who built the marketplace.

### Signals that separate staff from freelancers

1. **Location.** Look for Palo Alto, San Francisco, US remote and Berlin, with the Netherlands worth adding because of Lifted's acquisitions there. A senior "Upwork" engineer in one of those places is far more likely to be staff than one in a freelancer cluster such as Punjab (Pakistan), Dhaka or Lagos.
2. **Product names.** Staff mention the products the company is backing. Look for Uma (Upwork's AI work agent), the Upwork MCP server that lets AI agents act as clients, Upwork Now, Business Plus, and Lifted (the enterprise subsidiary).
3. **Elance and oDesk history.** A profile that shows Elance or oDesk before 2015 and then continues into Upwork is a strong sign of a long-tenured employee rather than a platform seller.
4. **Team vocabulary.** Matching engineers write about ranking, search relevance and recommendations. T&S engineers write about fraud, identity verification and risk scoring. Payments engineers write about payouts, KYC and cross-border money movement. A freelancer's headline sells services to clients. It rarely describes the platform's own systems.

This is tedious to do by hand with Boolean strings, because every product name also shows up in freelancer headlines. It's the kind of search [Refolk](https://www.refolk.ai/) was built for. You describe the person in plain English ("trust and safety engineers who worked at Upwork, based in the Bay Area or Berlin") and get back a ranked list based on the whole profile, not just the employer field.

> **Example search:** "Find senior engineers who worked on search, matching, fraud or payments at Upwork, Elance or oDesk, based in the San Francisco Bay Area, US remote, or Berlin, excluding freelancers who sell services on the platform." - [run it on Refolk](https://www.refolk.ai/s/ntj73yy9cr)
>
> A ranked shortlist of likely Upwork staff, with the signals that matched (team, product, location, tenure) shown on each profile.

## Are the 1,600 contractors worth recruiting too?

Yes. Upwork used about 1,600 independent contractors on internal projects during 2025, 2.5 times its employee headcount. They aren't part of the 37%, but many have verified experience building the marketplace.

This group is easy to overlook because their profiles look like freelancer profiles, which they technically are. The difference is the work. A contractor who spent real time on Upwork's matching or fraud systems carries the same domain knowledge as a full-time employee, and can usually start right away.

To find them, use the same product and team signals listed above, and then look for profiles where "Upwork" appears in the project history rather than as a storefront. A freelancer's headline sells services to clients. An internal contractor's profile describes systems they worked on.

## Why should Berlin be in your search?

Because the November round includes a separate reduction at an unnamed subsidiary, and it came right after Lifted, Upwork's enterprise subsidiary, bought WorkMotion, a Berlin employer-of-record (EOR) provider. An EOR provider legally employs workers in countries where the client has no entity of its own and handles their payroll and compliance.

Lifted had already bought Bubty in the Netherlands, along with Ascen. The filing doesn't say which subsidiary is cutting. My inference, which isn't confirmed, is that integration overlap in EOR, payments and compliance roles accounts for part of it, and some of that overlap is in the EU.

That's a different talent pool from May's. The first wave was about product and engineering team structure. This one may include people with European payroll, cross-border compliance and contractor-payments experience, which is useful to any fintech, HR-tech or marketplace company that pays people in more than one country. If you only search the Bay Area, you'll miss them.

## What did these engineers actually build?

They built ranking, fraud and payments systems for a marketplace that processed $4.0B in gross services volume in 2025.

It also matters why they're being let go. Upwork is still profitable. Q3 2026 GAAP net income was $31.8M, and the adjusted EBITDA margin was 34.4%. Revenue, on the other hand, fell 8.0% to $185.6M, and gross services volume fell 10.1% to $914.1M. Q4 revenue guidance is $169M to $177M.

> Upwork is cutting to protect its margins, not because these people underperformed. Read the profiles with that in mind.

Demand shrank and the company protected its margin. The people leaving weren't singled out for performance. That puts them in a similar position to the AI-cited peer cuts from May 2026, when Cloudflare cut 20% and Bill Holdings cut up to 30%.

The obvious rival employer for this talent is Fiverr, along with any other two-sided marketplace that needs matching and T&S experience. If you're recruiting marketplace matching engineers, assume someone else is already working through the same list.

> **Senior engineering roles open now** - [see the open roles](https://www.refolk.ai/jobs/search?family=engineering&seniority=senior)
>
> Current senior engineering openings, the level most of Upwork's marketplace staff held.

## When should you reach out?

Reach out now, before January. The restructuring isn't expected to be mostly done until Q1 2027, and it depends on "local-law requirements," so many affected people are probably still in notice periods, especially in the EU.

The timing works in your favor. Someone still in an EU notice period isn't looking hard yet, but they will be once January hiring picks up and every recruiter who saw the filing sends messages at once, competing marketplaces included. If you reach out in late November or December, you're one of a few messages, not one of dozens.

The hard part is building the list. Going through 59,156 headline matches to find a few hundred real staff, while also checking contractors and a subsidiary nobody has named, is slow manual work. Most teams don't have that time, so they either pay an agency or settle for whoever applies.

When you write to them, be specific. Name the system they worked on (matching, fraud, payouts) and the problem your team has at its scale. Don't open with the layoff. They already know about it, and recruiters who open with it tend to sound like they're circling a wreck. Opening with the work sounds like a job offer.

## FAQ

### How many people are affected by the Upwork layoffs in November 2026?

About 233, if you apply the 37% reduction to the roughly 630 employees reported in Upwork's FY2025 10-K. The May plan at 24% worked out to about 151, so the November expansion adds roughly 82 people. The 37% may include subsidiaries, so treat these figures as estimates rather than exact counts.

### Why can't I just filter LinkedIn by current company to find Upwork employees?

Because many freelancers list Upwork as their employer to show where their ratings and client history live. In Refolk's index, 59,156 profiles mention Upwork in their headline, about 94 for every real employee. Freelancer-titled profiles list Upwork as their employer at twice the rate of engineering-titled ones. Use location, product names such as Uma and Business Plus, and Elance or oDesk history instead.

### What roles should I look for if I want to hire laid-off Upwork engineers?

Focus on marketplace search and matching, trust and safety (fraud, identity, risk), and payments (payouts, KYC, cross-border). Because the subsidiary cut came right after Lifted bought Berlin-based WorkMotion, EOR, payroll and compliance engineers in Berlin and the Netherlands may also be in the pool, though the filing doesn't say which subsidiary is cutting.

### Are Upwork's internal contractors worth sourcing?

Yes. Upwork used about 1,600 independent contractors on internal projects in 2025, 2.5 times its employee headcount. They aren't counted in the layoff, but many have real marketplace engineering experience and can usually start right away.

---

_From the Refolk blog. More posts: https://www.refolk.ai/blog/index.md_
