Big Tech / FAANG Workforce, Updated May 2026

Big Tech Layoff Recovery (2026): The Engineer's Title-Translation and 90-Day Playbook

113,863 tech workers were cut in 2026, including roughly 16,000 at Amazon in Q1 alone, 20,000 to 30,000 at Oracle, 8,750 at Microsoft, and 8,000 at Meta. Your level codes (L5, E6, MSFT 65) do not survive the trip to a startup ATS, and your H-1B clock starts at your last paid day, not your severance end. This guide gives you the leveling crosswalk, the severance math, the promo-doc-to-plain-English translation, and a 90-day plan calibrated to where ex-FAANG engineers are actually landing.

By the numbers (sourced)

Find out how your L5, E6, or MSFT 65 reads to a Series B startup ATS

Upload your FAANG resume and a target job description. Get your ATS score, missing keywords, and a leveling-translated rewrite in 30 seconds. Free.

Scan your resume free

The 2026 layoff map and why this round is different

The 2026 cuts are not a repeat of 2023. The post-ZIRP correction is mostly absorbed; what is happening now is AI-driven org consolidation and a capex shift from headcount to GPU clusters. Amazon eliminated roughly 16,000 corporate roles in Q1 2026, Oracle is in the middle of a 20,000 to 30,000 reduction, Microsoft cut around 8,750, Meta cut around 8,000, and Google continues to run rolling waves through the cloud, ads, and hardware orgs. TrueUp is tracking 113,863 impacted tech workers across roughly 280 companies year to date.

The composition is also different. The 2023 round was concentrated in recruiting, support, and ops. The 2026 round is hitting senior engineers, middle managers, and entire mid-tier PM functions, which means the supply of L5 to L7 ICs on the market is unusually high. That is the bad news. The good news is that the AI labs (Anthropic, OpenAI, Scale, Databricks) and a small set of cleared infra companies (Anduril, Palantir) are hiring exactly that profile, often at flat or higher total comp.

The job you are competing for is no longer your old job at a different ticker. It is a smaller, leaner team where your scope label ("Staff at Google") is read with skepticism until you prove it. The rest of this page is about closing that gap on paper, in the ATS, and in the first 90 days after your last paid day.

Title and level translation: Amazon, Google, Meta, Microsoft to market

Recruiters at Series B to D startups and AI labs do not parse L6 or E5 the way an internal calibration committee does. Use this crosswalk to set the title on your resume header, then keep the original level in parentheses only if you are applying to another big tech company.

AmazonGoogleMetaMicrosoftMarket titleYears exp
L4 (SDE I)L3 (SWE II)E3 (SWE)59-60Software Engineer (Junior)0-2
L5 (SDE II)L4 (SWE III)E4 (SWE)61-62Software Engineer (Mid)2-5
L5 senior-light / L6 (SDE III)L5 (Senior SWE)E5 (Senior SWE)63-64 (Senior SDE)Senior Software Engineer (or Staff at small Series B)5-9
L6 (Sr SDE / Sr Mgr)L6 (Staff SWE)E6 (Staff SWE)65 (Principal SDE)Staff or Principal Engineer9-13
L7 (Principal)L7 (Sr Staff SWE)E7 (Principal SWE)66 (Partner)Principal Engineer or Director13-18
L8 (Sr Principal / Director)L8 (Principal SWE)E8+ (Distinguished)67+ (Distinguished / Technical Fellow)Distinguished Engineer or VP Engineering18+

Calibration varies. A Google L5 with 7 years tenure often maps cleanly to Senior at a Series C startup but to Staff at a 50-person Series B. A Meta E6 typically lands as Staff at AI labs and as Principal at enterprise SaaS. Compensation source and level mapping cross-checked against Levels.fyi.

Translate your level codes to a startup the same way you translated this table

Drop in a Series B or AI-lab job description. We map your Amazon L6 / Google L5 / Meta E5 / MSFT 65 promo-doc bullets into language a 200-person startup ATS understands.

Translate my levels

Severance terms compared

These are the published or widely reported severance frameworks for 2026 rounds. Always verify against your own separation agreement; numbers below are typical, not guaranteed, and individual offers vary by tenure, geography, and stock plan.

CompanyBase weeksPer-year addCapCOBRA subsidyRSU acceleration
Amazon (corporate)~2 weeks per year of service (no flat base)Tenure-based3 to 26 weeks typical capUp to 6 monthsPro-rated to next vest only
Google16 weeks+2 weeks per yearNo published hard capUp to 6 months60 days of additional vest
Meta16 weeks+2 weeks per yearNo published hard capUp to 6 monthsPro-rated; partial acceleration
MicrosoftVaries by program (tenured separation 6+ months for long-tenure cases)Tenure-basedProgram-specificUp to 6 monthsPro-rated to next vest
OracleVaries; reported 4 to 8 weeks base for many 2026 casesLimited per-year addProgram-specificLimitedGenerally none

Severance comparison aggregated from SeveranceCalc 2026 reporting and individual employee disclosures. Always read your own separation agreement; signing it typically waives your right to negotiate later.

The H-1B 60-day grace clock starts at your last paid day, not severance end

USCIS guidance is explicit: the 60-day H-1B grace period after a nonimmigrant termination runs from the last day of paid employment, not from the end of your severance period. If you are terminated on March 1 with 16 weeks of severance through June 21, your grace period still ends on April 30. You must file a change of status, transfer your H-1B to a new employer, or depart the US by that date. Do not rely on severance pay extending status.

Source: USCIS, "H-1B Specialty Occupations" guidance on options after termination

Promo-doc and internal jargon to translate for non-FAANG resumes

Every term on the left is calibration-committee language. Outside of big tech it sounds either generic or actively confusing. Replace each with its market translation when applying to startups, AI labs, or non-tech enterprises.

Scope
Internal: blast radius of your work. Translate to: team size, system criticality, dollar impact, or user count.
Impact
Internal: the metric you moved at promo. Translate to: revenue lifted, latency cut, cost avoided, or users acquired (always with a number).
Ownership
Amazon LP language. Translate to: "end-to-end responsibility for the X service across design, on-call, and roadmap." Avoid quoting the LP by name.
OE (Operational Excellence)
Amazon-internal pillar. Translate to: reliability work, SLO definition, incident response, or cost-to-serve reduction.
Customer obsession / dive deep
Amazon LPs. Translate to specific behaviors: "ran weekly customer interviews" or "audited 200 production traces to root-cause a regression."
OKRs
Google-origin planning artifact. Most startups understand this verbatim; in non-tech enterprise translate to "quarterly goals" or "performance objectives."
Bar Raiser
Amazon hiring role. On a resume, translate to: "trained interviewer for senior IC roles" or "hiring committee member, 200+ loops."
Tier-1 Service / Tier-0 dependency
Internal reliability tier. Translate to: "customer-facing service with five-nines SLO" or "foundational platform serving N internal teams."
On-call rotation
Keep verbatim; this is universal in software. Add the rotation cadence (e.g., "primary on-call, 1 week in 6") for clarity.
Design doc / RFC
Keep verbatim. Most engineering-led startups use the same vocabulary; some prefer "technical proposal" or "ADR."
Code review velocity / CR throughput
Internal productivity metric. Translate to: "reviewed 30+ pull requests per week across 4 services" or omit if not load-bearing.
Promo-doc voice (third person, ghostwritten)
Replace with first-person, active-voice bullets. "Drove a 22% latency reduction" beats "Was responsible for driving a 22% reduction in P99 latency."
Scope-creep / scope-grow
Internal calibration phrasing. Translate to: "expanded charter from a single service to a 4-team platform."
L-band / level band
Calibration shorthand. Delete from the resume body; map to a single market title in the header.

Where FAANG talent is actually landing in 2026

These are the ten employers absorbing the largest share of impacted FAANG, Amazon, and Microsoft engineers in 2026 based on TrueUp post-layoff destinations and public hiring announcements. The Series B to D AI labs are the hottest market; cleared employers like Palantir and Anduril preferentially hire ex-AWS and ex-Google security and infra engineers.

Anthropic

Research (San Francisco), Applied (San Francisco, Seattle, NYC, London)

Highest demand: ex-Google research engineers and ex-Meta infra. Mostly L5 to L7 equivalent; titles flatten to Member of Technical Staff. Compensation is competitive with Meta E6 plus equity upside.

OpenAI

Research, Applied, Platform (San Francisco, NYC, Dublin)

Hires aggressively from Google DeepMind, Meta FAIR, and AWS infra. Distributed systems engineers (Spanner, BigTable, RDS lineage) are in demand for inference scaling.

Databricks

San Francisco, Seattle, Mountain View, Amsterdam

Strong fit for ex-AWS, ex-Snowflake, and ex-Google data infra. Maps Amazon L6 cleanly to Staff Software Engineer; recognizes Google L5 as Senior.

Stripe

South San Francisco, Seattle, NYC, Dublin (remote-friendly for senior engineers)

Hires senior backend, payments infra, and ML risk engineers. Calibration is rigorous: Google L5 maps to L4 at Stripe, L6 to L5; expect a downlevel discussion.

Cloudflare

Austin, San Francisco, London, Lisbon

Heavy hiring of ex-AWS network and edge engineers. Title compression is real: Amazon L6 often lands as Senior, not Staff. Total comp gap closed by faster RSU vesting.

Scale AI

San Francisco, St. Louis (Defense), DC (Federal)

Defense and government cluster hires ex-AWS GovCloud, ex-Google Federal, and TS/SCI cleared engineers. Commercial side hires standard ex-FAANG ML and platform engineers.

Palantir

Denver, NYC, DC, Seattle, Palo Alto

Preferential pipeline for ex-AWS, ex-Google security, and any cleared engineer. Compensation jump for cleared FAANG engineers is typically 20 to 40%; recognizes L6/E6 as Forward Deployed Engineer Lead or Tech Lead.

Anduril

Costa Mesa CA, Atlanta, DC, Seattle, Boston (Lattice and Roadrunner)

Hires ex-AWS, ex-Google, ex-Meta engineers with US person status; cleared candidates preferred for Roadrunner and Bolt programs. Maps L6/E6 to Senior or Staff Engineer with hardware-software systems exposure.

MongoDB

NYC, Austin, Palo Alto, Dublin

Targets ex-AWS DynamoDB and ex-Google Spanner engineers for Atlas Search and Vector Search teams. Calibration close to Stripe's; expect a downlevel of one band.

Snowflake

San Mateo, Bellevue, Berlin

Heavy hiring of ex-Databricks, ex-AWS Redshift, and ex-Google BigQuery engineers. Titles map cleanly to Snowflake's L1-L7 IC ladder; comp is competitive with Meta E6.

Pick a destination employer, run your resume against one of their open roles

Anthropic, OpenAI, Databricks, Stripe, Anduril, Palantir, Scale AI. Tailor in one session. Beat their internal ATS with the right scope-and-impact rewrite.

Match my resume to a hiring company

Before and after: an Amazon L6 SDE bullet rewritten for a startup ATS

Same accomplishment. The before version reads as a promo doc excerpt with internal jargon and LP voice. The after version translates each term to startup vocabulary and leads with quantified outcome.

Before
Demonstrated Ownership and Customer Obsession by driving Operational Excellence improvements across a Tier-1 service, partnering with 3 sister teams to dive deep on root causes of P99 latency regressions and raising the bar for our on-call rotation through process and tooling investments.
After
Cut P99 latency 38% on a customer-facing payments service (12B requests/month) by re-architecting the read path and rolling out a new circuit-breaker library across 4 partner teams; reduced primary-on-call pages by 60% in the same quarter.

The before version names three Amazon LPs (Ownership, Customer Obsession, Operational Excellence), uses the internal Tier-1 classification, and ghostwrites in calibration-committee voice. A startup recruiter recognizes none of that as signal. The after version keeps the technical scope (P99, payments, 12B requests, circuit breaker) and adds the two numbers that matter at any company: 38% latency cut and 60% pager reduction.

What to remove from your big-tech resume

These five patterns mark a resume as ex-FAANG to a startup recruiter and they all cost you. Strip each before you start applying.

Don't: Cryptic internal team names ("AWS PXT Velocity Tier-2", "GCP IBP Sigma", "Meta WIT Backend").

Do: A short scope description ("AWS internal HR platform serving 1.5M employees", "GCP internal billing pipeline, $45B ARR").

Why: Internal team names mean nothing outside the company. Recruiters bounce on them in 6 seconds.

Don't: Internal tool names without context ("Apollo", "Brazil", "Spanner UI", "Hack", "Phabricator").

Do: The category, then the internal name in parentheses if it adds signal ("internal CI/CD platform (Apollo)", "build system (Brazil, Java/Python)").

Why: Some internal tools are recognizable (Spanner, Hack); most are not. Lead with the category so the ATS can match against industry keywords.

Don't: Level codes embedded in the title ("Senior SDE II (L6)", "Staff Software Engineer L7", "E5 SWE").

Do: A clean market title ("Senior Software Engineer", "Staff Software Engineer") with the level code optional, in parentheses, only when applying to another big tech.

Why: Level codes signal calibration overconfidence to startups and they do not parse correctly in most ATS. Workday and Greenhouse will index "L6" as a literal string, not as a seniority signal.

Don't: Promo-doc voice (third person, hedged, packed with LP names and internal pillar references).

Do: First-person, active-voice bullets that lead with an action verb and end with a number.

Why: Recruiters skim bullets in 6 seconds. Promo-doc voice triples reading time and reduces the number of bullets they actually scan.

Don't: Overuse of LP names ("Bias for Action", "Earn Trust", "Have Backbone, Disagree and Commit").

Do: The behavior the LP describes, expressed in plain English ("shipped a v1 in 2 weeks against engineering pushback; data validated the call within 30 days").

Why: LP names only mean something inside Amazon. Outside, they look like jargon and they crowd out the actual achievement.

ATS keywords that travel from FAANG to startup or AI lab

These are the 24 cross-company keywords that translate cleanly from a big tech resume into a startup, AI lab, or scaleup ATS. Add them only where they reflect actual work; never keyword-stuff.

Distributed systems

Microservices

Kubernetes

gRPC

Kafka

Spark

Airflow

Terraform

AWS (EC2, S3, DynamoDB, Lambda)

GCP (BigQuery, GKE, Spanner)

Python

Go

Rust

TypeScript

PyTorch

JAX

LLM inference

Vector databases

RAG pipelines

Observability (Datadog, Grafana, OpenTelemetry)

SLO and SLI definition

Incident response (PagerDuty)

Code review and design review

Mentorship (mentored N engineers)

All of these keywords. One scan. Before your H-1B clock expires.

Free first scan, no signup. We score against the keyword grid above and any startup JD you paste.

Run my free scan

Your 90-day post-layoff plan with H-1B and severance math

This timeline assumes you have your separation date in hand. If you are H-1B, anchor every milestone to your last paid day, not your severance end. If you are a US citizen or permanent resident, you have more runway but the same job market.

  1. 01

    Days 1-3: Pull records, lock the H-1B math, and read the agreement

    Download your performance history, last 6 months of design docs, and any patents or publications from your corporate drive before access is revoked. If you are H-1B, calculate your day-60 deadline from your last paid day (not severance end) and write it down. Read your separation agreement before signing; once signed, your right to negotiate is typically waived.

  2. 02

    Days 4-7: Decide whether to negotiate severance

    Most big-tech severance offers are negotiable in narrow ways: extended COBRA, an additional RSU vest cliff, or 2 to 4 additional weeks for unique circumstances (medical, immigration, parental). Negotiate before signing and in writing. If you have an offer letter discrepancy or unpaid bonus, raise it explicitly; companies often resolve those without further negotiation.

  3. 03

    Days 8-14: Translate the resume from promo-doc to startup voice

    Cut every LP reference, internal tool name without context, and level code from the title bar. Replace the calibration-committee voice with first-person active-voice bullets. Map your level to a single market title using the crosswalk above. Save as PDF, plain sans-serif, no columns.

  4. 04

    Days 15-21: Build the LinkedIn rewrite and turn on Open to Work (private)

    Title bar should read your translated market title (Senior or Staff Software Engineer), not your old level. About section should open with the macro context ("10 years building distributed systems at Amazon and Google, now looking for staff IC roles at AI infra companies"). Turn on Open to Work in private mode (recruiters only). Update the About section before you update the headline; LinkedIn search ranks on About content.

  5. 05

    Days 22-35: Apply to 30 to 50 roles, batched by employer cluster

    Use the employer map: 8 to 10 AI labs, 8 to 10 cleared infra (if you have a clearance or US person status), 8 to 10 enterprise SaaS (Stripe, Snowflake, MongoDB, Cloudflare), and 6 to 8 mid-stage startups. Tailor only the top third of the resume per role. Use ResumeAdapter to scan against each JD before submitting.

  6. 06

    Days 36-50: Reactivate the network for referrals

    Reach out to 25 to 40 ex-coworkers who already left for the companies on your shortlist. Ask each for one referral, not a job. Referral conversion at AI labs and cleared employers is 4 to 6x cold application. Keep a single spreadsheet: company, contact, date asked, response, follow-up date.

  7. 07

    Days 51-70: First-round loops, leveling pushback, and comp benchmarks

    Expect downlevel offers from at least 30% of the companies you interview at; this is normal, not personal. Use Levels.fyi to benchmark before negotiating. If you are H-1B, prioritize companies with documented H-1B transfer support and premium-processing willingness; ask explicitly in the recruiter call.

  8. 08

    Days 71-90: Close one offer, validate against severance and immigration constraints

    If you are H-1B and your day-60 is approaching, file for change of status or transfer immediately on offer signature; do not wait for the start date. Run any final offer through a comp benchmark (Levels.fyi, Blind, Glassdoor) and an immigration lawyer if your status is changing. Decline gracefully on offers you do not take; that network compounds in the next cycle.

FAQ

Frequently asked questions

Sources cited in this guide

  1. [1]
    Computerworld 2026: "Tech layoffs this year, a timeline"

    Primary source for the Oracle 20,000-30,000 figure and Microsoft 8,750.

  2. [2]
    TrueUp Layoffs Tracker

    Source for the 113,863 total impacted tech workers in 2026.

  3. [3]
    Crunchbase News: Tech Layoffs

    Source for FAANG alumni search-length data and post-layoff destination tracking.

  4. [4]
    Levels.fyi (leveling and compensation database)

    Authoritative source for the Amazon, Google, Meta, Microsoft level crosswalk and comp benchmarks.

  5. [5]
    USCIS: H-1B Specialty Occupations (60-day grace period guidance)

    Authoritative source for the 60-day H-1B grace clock starting at last paid day.

  6. [6]
    Newsweek (2026): "All the tech giants announcing sweeping layoffs in 2026"

    Source for the Amazon ~16,000 Q1 2026 figure and Meta ~8,000.

  7. [7]
    SeveranceCalc: 2026 Tech Layoffs Severance Comparison

    Source for Google and Meta severance formulas (16 weeks + 2 weeks per year of service).

Stop guessing.
Scan your resume.

Upload your resume and a target job description. Get your ATS score, missing keywords, and a structured rewrite in 30 seconds.

Scan your resume free