TikTok's Aug 19 Bellevue WARN: 75 Cut, 37 Ranking-ML Profiles in Play

TikTok filed a WARN notice in Washington on August 19, 2026, cutting 75 e-commerce roles at Lincoln Square North in Bellevue with an October 19 separation date. Every recruiter in Seattle is about to chase the same short list, and most will search under the wrong company name and miss half of it.

The people who move first, and who search by the right entity, get the ranking ML engineers before Meta Shops and Amazon Rufus quietly close them.

What the WARN filing actually says

TikTok is cutting 75 workers in the Seattle region, filed under a corporate entity most sourcers will miss. The notice, submitted by TT Commerce & Global Services LLC on TikTok letterhead with two ByteDance employees listed as contacts, targets almost entirely TikTok Shop and Global E-Commerce roles at 700 Bellevue Way NE.

The named job families in the filing:

• Anti-fraud and governance program managers
• Seller and creator operations staff
• Campaign managers
• Data scientists
• Backend and frontend engineers

Separations are scheduled for October 19, 2026, with at least 60 days' notice. Affected employees can apply internally first; those who do not land a new role are separated on that date. That is the source of the "30-day external window" that gets thrown around, and, as I will argue below, that window is a myth. The urgency is real, but the shape is wrong.

For sourcing purposes, three details matter more than the headline number:

1. The filer entity is TT Commerce & Global Services LLC, not "TikTok." LinkedIn keyword searches that only match "TikTok" will miss profiles that list the LLC or "ByteDance."
2. The site is Lincoln Square North, Bellevue (700 Bellevue Way NE). Radius filtering works.
3. This is the second Bellevue cut in 24 months. 65 employees were laid off in 2025, bringing cumulative Bellevue reductions to 140.

Why the "340 pool" framing is wrong

The real headline number is 75 in Bellevue, not 340. The 75 plus 250 from the Nashville joint-venture office closure on October 5, 2026 gets you to ~325. That is a fair number for the total displaced US TikTok workforce this month. It is not the number of ranking ML engineers you can hire.

The WARN filing spans product, ops, fraud, campaigns, and engineering. The data scientist plus software engineer subset with ranking or recommendation systems experience is closer to 20 to 35 people, not hundreds. Selling "340" to a hiring manager sets an expectation the market cannot meet. Sell specificity instead.

37

The real pool: 37 people, and Meta already has 8 of them

In Refolk's index of professional profiles, only ~37 US profiles match the intersection of (Data Scientist OR ML Engineer) AND (Recommendation Systems OR Ranking). Widen the search to include Software Engineers and Applied Scientists with recsys skills and you still only get ~35. This is the entire competitive set that will pursue displaced TikTok Shop rankers.

Where those 37 people work today, and where they live, decides who wins the next 45 days.

Meta employs 22% of this pool. Reddit has 3, Apple has 2, and Google and Scale AI round out the top employers. By region, the top three are SF Bay Area (5), Mountain View (5), and Seattle (4). Seattle has four people. Four.

The displaced TikTok Shop cohort will roughly double Seattle's local recsys DS/MLE bench overnight. That is a one-time supply shock. You will not see it again this cycle.

Why Meta absorbs ex-TikTok fast even after cutting 699 in Bellevue

Meta reduced its Bellevue workforce by 699 in May 2026, yet Meta still employs the largest share of the US recsys ML pool. The mechanism to understand is rebalancing, not headcount. Meta cuts and hires simultaneously because Reels and Shops ranking teams have specific reqs that stay open through org-wide RIFs. A closed data-labeling team in Bellevue does not close a Reels ranking req in Menlo Park or a Shops relevance req in Seattle.

The same logic applies to Amazon Rufus and Amazon Ads ranking, which are Seattle-local and do not require relocating ex-TikTok Bellevue staff. The specific ranking teams inside Rufus and Ads have narrow, senior reqs and can absorb three to five ex-TikTok Shop rankers without a headline.

That is the buyer set you are competing with:

• Meta Shops and Reels ranking (8 in the pool, national reach)
• Amazon Rufus and Amazon Ads ranking (Seattle-local, no relocation friction)
• Reddit (3 in the pool, aggressive post-IPO)
• Apple, Google, Scale AI (2, 1, 1 respectively)

If you are not one of these five, your pitch has to compensate for a slower brand-recognition cycle. Get to inbox on day 1.

The 30-day window is a myth. Reach out day 1.

The WARN notice gives 60 days and an internal-transfer option, so most articles will tell you the external window is 30 days. In practice the best ML people start externally interviewing in week 1, because internal ByteDance roles for e-commerce ranking are thin after the January 22, 2026 USDS Joint Venture restructure with Oracle, Silver Lake, and MGX.

The internal-transfer option is real for program managers and ops. For ranking ML, the internal market is smaller than the external one. Reach out the first week the WARN goes public. That is right now.

The internal-transfer option is real for program managers. For ranking ML, the external market is bigger than the internal one.

A day-1 message that lands well tends to have three properties:

1. It names TikTok Shop or Global E-Commerce specifically, not "TikTok."
2. It references a nearby competing offer set (Meta Shops, Rufus) without pretending you are them.
3. It respects that the recipient has 60 days of pay and is not desperate.

The generic "saw the news, would love to chat" opener from Bay Area recruiters gets ignored. A Bellevue-based startup founder asking about a Shops-relevance eval methodology gets a reply.

How to search: the entity, the address, and the Blind thread

Sourcing this cohort requires five keyword filters, one address radius, and one active forum. Miss any of them and you leave half the pool on the table.

The five keyword filters to OR together on LinkedIn:

• TikTok
• TikTok Shop
• Global E-Commerce (the internal division name)
• TT Commerce & Global Services (the WARN filer entity)
• ByteDance (the parent, listed as the contact employer)

The address radius: 700 Bellevue Way NE, Bellevue, WA. Set a 5-mile radius on LinkedIn's location filter to catch employees who list Bellevue, Redmond, or Kirkland.

The forum: the TikTok thread on Blind (teamblind.com/company/TikTok/posts/tiktok-layoffs). A poster on August 19 confirmed "Big layoffs today at Bytedance, across GNE, Ecommerce, Customer service and others. Mostly non mainland chinese employees." That thread is a live signal of who is impacted and open to outreach. Cross-reference names surfaced there against LinkedIn's Bellevue radius.

This is exactly the sort of multi-signal search that eats an afternoon in LinkedIn Recruiter and takes 90 seconds in Refolk. Ask in plain English for "TikTok Shop or Global E-Commerce data scientists and ML engineers in the Seattle area with ranking or recsys experience" and you get the intersection across LinkedIn, GitHub, and the open web, ranked, with the entity aliases already resolved. That last part matters more than the ranking; most tools miss the LLC name.

The Seattle-market context that shrinks your local competition

Adjacent August 2026 Seattle-area cuts drain the same recruiter market you are working, which is bad for the displaced but good for the sourcers who move now. Local hiring managers are distracted by their own RIFs.

The same-month Seattle-area context:

• Zillow cut more than 500 jobs in August 2026, including 91 in Washington state
• Microsoft eliminated 605 positions in the state in July as part of a broader 4,800 reduction
• Google cut 52 jobs locally in August
• Salesforce cut 59 locally in August
• Bungie cut 292 at its downtown Bellevue HQ in June 2026
• Meta reduced Bellevue by 699 in May 2026

Two implications. First, the Bellevue market can absorb and shed talent fast; do not assume ex-TikTok Shop rankers stay put waiting for the perfect offer. Second, several of your usual competitors have distracted or paused local ranking-team recruiters this quarter. If you are a Series B ecommerce startup with a ranking problem, this is the widest sourcing window you will get against the incumbents in a year.

FAQ

How many ex-TikTok data scientists and engineers will actually be on the market?

The August 19, 2026 WARN filing lists 75 total roles in Bellevue across product, ops, fraud, campaigns, and engineering. The data scientist and software engineer subset with ranking or recommendation systems experience is roughly 20 to 35 people. Add the 250-person Nashville office closure separating on October 5, 2026 and you get ~325 total displaced US TikTok employees this quarter, but the ranking ML subset stays small.

Which employers will absorb this talent fastest?

Meta Shops and Reels ranking, Amazon Rufus and Amazon Ads ranking, and Reddit. In Refolk's index, Meta already employs 22% (8 of 37) of the US recsys DS/MLE pool, which is why they absorb ex-TikTok fast even after cutting 699 in Bellevue in May 2026. Amazon competes hardest in Seattle because there is no relocation friction from Bellevue.

What LinkedIn keywords should I search for?

Use OR filters across all five names: TikTok, TikTok Shop, Global E-Commerce, TT Commerce & Global Services, and ByteDance. The WARN filing was submitted by TT Commerce & Global Services LLC, so employees listing the LLC as their employer will not appear in a plain "TikTok" search. Radius-filter on 700 Bellevue Way NE.

Is the 30-day sourcing window real?

Kind of. The formal internal-transfer option runs the full 60 days to the October 19, 2026 separation, but the most sought-after ranking ML engineers start externally interviewing in week 1 because internal ByteDance ecommerce roles are thin post-USDS JV restructure. Reach out on day 1, not day 30.