Redis Just Freed 80 Tel Aviv Engineers Into a Senior Pool of 91
Redis cut 80 Tel Aviv R&D engineers on July 13, 2026. Israel's senior AI infra pool is only 91. Here's how to source the agent-memory builders first.
On July 13, 2026, Redis laid off about 80 of its 300 Tel Aviv R&D engineers, roughly 27% of the local workforce, at a company that just crossed $300M ARR and signed a ten-year lease down the street. The same week, Together AI closed $800M and Chai Discovery closed $400M, both in categories that live or die on the exact retrieval, vector, and agent memory skills these engineers spent the last year building. If you run sourcing at an AI infrastructure startup, this is the shortest window you will get all year.
Why Redis cut 80 engineers while ARR was still growing
This was a strategic reshuffle, not a distressed layoff, and that changes who is actually on the market. Redis passed $300M ARR "driven by increasing demand for AI infrastructure and as agentic AI systems grow," per its own statement to Globes, and it signed a fresh decade-long Tel Aviv lease weeks before the cut. The mechanism matters: when a growth-stage infra company reshuffles, the engineers on the old stack get released, and the AI-context team gets protected.
That is exactly the recruiter's problem. The 80 people walking out include both:
- Classic in-memory KV specialists who kept the legacy Redis Enterprise business running.
- A smaller, more valuable cohort who shipped Redis's new context and memory platform for AI agents earlier this year.
Both groups are hireable. Only the second group is the prize. The first question in any intro call is which team the candidate was on, and which service they owned in the last two quarters. If they cannot name a memory-platform component, they are a strong backend hire, not an agent-infra hire.
The broader Israeli context also matters. Wix, Rapyd, and Amdocs all announced cuts in the same window, and the strong shekel is making local operations more expensive for dollar-priced buyers. Supply is spiking. Prices are not, yet.
What the Refolk index says about the addressable pool
There are 583 Israel-based engineers with Redis or vector database skills, but only 91 senior+ people in Israel with real LLM, RAG, or vector-search experience. That second number is what makes this layoff a market-moving event.
In Refolk's index of professional profiles, the Israeli talent map for this skill set breaks down like this:
| Segment | Count | Note |
|---|---|---|
| Israel engineers with Redis or vector DB skills (SWE / backend / infra) | 583 | Base pool |
| Israel senior+ with RAG, LLM, or vector-search skills | 91 | The scarce tier |
| Redis Tel Aviv R&D headcount pre-cut | 300 | Times of Israel |
| Redis Tel Aviv laid off July 13 | ~80 | Globes |
| Layoffs as % of the scarce senior AI-infra pool | ~88% | 80 / 91 |
| Layoffs as % of total Israeli Redis / vector-DB engineer pool | ~14% | 80 / 583 |
Read that fifth row twice. Even if only half of the departing 80 are senior, this single event just added roughly 40 senior AI-infra engineers to a national pool of 91. That is a step change in liquidity, and the market has not priced it yet because the names are not on LinkedIn's "Open to Work" filter and will not be for another two to six weeks.
Agent memory is not vector DB, and most sourcers get this wrong
"Knows Pinecone" is not the same skill as "shipped agent context infra," and confusing the two is why most Boolean searches for this pool return the wrong 400 people. Agent memory platforms coordinate state: session history, tool call traces, business data retrieval, and the eviction and summarization logic that keeps a running agent from blowing its context window. Vector search is one component. Retrieval quality, cache hierarchy, and freshness guarantees are the harder ones.
The Redis memory platform, per the company, "searches, gathers, and coordinates business data that AI agents need to perform and complete personalized tasks fast and accurately." Translate that into resume signal:
- Owned or contributed to a production semantic cache or retrieval layer sitting in front of an LLM.
- Shipped tool-call state or session memory for multi-step agents.
- Wrote about vector index tuning, hybrid search, or memory eviction policies.
- Handled the latency budget between agent turns (single-digit millisecond retrieval, not batch).
Fewer than a fraction of the 583-strong Israeli Redis and vector talent pool has actually done this. That fraction is now heavily concentrated in the outgoing 80. Describing the profile in plain English is the exact gap Refolk closes: instead of chaining Boolean operators against LinkedIn skill tags that everyone claims, you ask for "Tel Aviv engineers who shipped agent memory or semantic caching in the last 18 months" and get a ranked shortlist.
Who is competing for these 80 engineers, and how fast
Assume Together AI, Decart AI, monday.com, Gong, and Yiftach Shoolman's new startup are all in the market by end of month, and price accordingly. This is not speculation; the funding and hiring signals are already on the tape.
The competitive set, ranked by how aggressive they will be:
- Together AI. Closed an $800M Series C at $8.3B on July 1, 2026, led by Aramco Ventures with NVIDIA and General Catalyst. Plans to expand infrastructure footprint 50x over five years. Needs inference and retrieval engineers immediately. Will make dollar-denominated remote offers into Tel Aviv, which the strong shekel makes attractive to both sides.
- Yiftach Shoolman's new startup. The Redis co-founder stepped back from day-to-day management last year and founded a new (undisclosed) company. He is a natural magnet for Redis alumni who liked the mission but not the reshuffle. Assume he has already texted the top 15.
- Decart AI. Israeli generative AI infra startup, already employs Redis and vector-skilled engineers per the index. Local, in-person, no relocation friction.
- monday.com and Gong. The two largest Israeli scale-ups already staffed with this profile. First call for anyone who wants to stay local, keep a similar comp band, and stop worrying about runway.
- Chai Discovery. Raised $400M on July 14, 2026, valuing the company at $3.8B, nearly triple its $1.3B mark from seven months earlier. SF-based, but hiring remote AI infra.
Israel's senior AI infra pool is 91 people. Redis just handed the market 80 new candidates in a single afternoon.
The pattern to notice: three of these five buyers are non-Israeli, well-funded in USD, and comfortable with remote. If you are recruiting locally, you are not competing against other Tel Aviv startups. You are competing against Together AI's comp band.
The 30-day playbook for sourcing Redis Tel Aviv alumni
Move in the next 30 days or lose the top third of this cohort to remote US offers. The sequence that works, in order:
- Day 0 to 3: Build the named list. Pull every current Redis Tel Aviv engineer, filter by tenure of at least 12 months, and cross-reference with contributions to the memory platform (blog posts, conference talks, GitHub activity on retrieval or memory repos). Aim for a list of 80 to 120 names.
- Day 3 to 7: Segment by project. Split into "AI context / memory platform" vs. "classic Redis Enterprise." Only the first segment gets the premium outreach template. The second gets a standard backend pitch.
- Day 7 to 14: First-touch outreach. Reference the specific project, not the layoff. "I saw you presented on session memory for agents at [X]" beats "sorry to hear about the news" by an order of magnitude in reply rate.
- Day 14 to 30: Close on speed, not comp. Israeli candidates in this pool will get three to five offers. Winning ones are decided on interview loop length (target: 8 business days end to end) and a specific technical mandate the candidate can visualize owning.
- Day 30+: Track the diaspora. Some will take three months off. Set reminders for October. Some will land at Shoolman's new company or Decart, then become recruitable again in 18 months when equity vesting cliffs hit.
The failure mode here is spending week one on Boolean queries against LinkedIn skill tags. "Redis" as a skill tag surfaces every backend engineer who ever ran a cache. That signal-to-noise ratio is the reason sourcing tools priced on volume lose to intent-priced tools on niche talent hunts. Describing the person in a sentence and letting Refolk rank across GitHub commits, LinkedIn tenure, and company blog bylines is the cheapest way to compress week one into an afternoon.
What this means beyond Redis: Israeli tech layoffs 2026
Expect two to four more mid-cap Israeli tech companies to announce similar reshuffles by end of Q3 2026, and start pre-building your bench now. The macro drivers, per Times of Israel and Calcalist reporting, are:
- The strong shekel making Israeli engineering labor expensive in dollar terms.
- Automation pressure (Wix, Rapyd, Amdocs) forcing headcount rebalancing away from legacy stacks toward AI-native teams.
- USD-funded acquirers finding remote Tel Aviv talent cheaper than SF or NYC hires with equivalent skills.
Israeli AI startups that treat each layoff as a one-off event will lose. The ones that maintain a warm bench of the 91-person senior pool, refresh it monthly, and know within 48 hours when any of them go "quietly open," will absorb this wave and the next two.
The Redis Tel Aviv event is the loudest signal, not the only one. If you are hiring vector database engineers or agent memory engineers in the second half of 2026, the batches worth watching are Redis (this month), whichever mid-cap announces next month, and the quieter Wix and Amdocs cuts already in progress.
FAQ
How many Redis Tel Aviv engineers were laid off on July 13, 2026?
Redis laid off approximately 80 employees out of 300 in its Tel Aviv R&D center, or about 27% of the local workforce, per Globes and Times of Israel reporting on July 13, 2026. Globally, Redis employs roughly 1,300 people, so the cuts were disproportionately Israeli. The company had just crossed $300M ARR and signed a new ten-year Tel Aviv lease, framing this as a strategic reshuffle rather than a distressed layoff.
How do I tell which Redis engineers worked on the agent memory platform?
Ask which internal service or product surface they owned in the last two quarters, and cross-reference with public artifacts. Engineers who built the memory platform will reference session state, tool-call coordination, retrieval latency budgets, or business-data gathering for agents. Engineers on the classic stack will reference cluster ops, replication, or Redis Enterprise features. Blog bylines, conference talks (Redis Days, KubeCon, AI Engineer Summit), and GitHub activity on retrieval-adjacent repos are the strongest external signals.
Who is most likely to hire the departing Redis engineers first?
Together AI, Yiftach Shoolman's new startup, and Decart AI are the three fastest-moving buyers, with monday.com and Gong close behind for candidates who want to stay local. Together AI closed $800M twelve days before the cut and needs exactly this skill profile. Shoolman, Redis's co-founder, is now running a new venture and is a natural landing pad for alumni. Assume the top 15 engineers will have three to five offers within two weeks of exit.
Is this pool worth pursuing if I am hiring outside Israel?
Yes, especially if you can offer remote work with a compelling technical mandate. The strong shekel makes Israeli senior engineers relatively cheaper for USD-funded acquirers, and the senior tier of this pool (91 people nationally before the layoff) is deep on production retrieval and memory experience that is genuinely scarce in the US and EU markets. The tradeoff is timezone overlap and the fact that local Israeli employers will compete hard on speed and cultural fit.