Meta Data Scientist Career Ladder
Every level of Meta's data science ladder from IC3 to IC6 — typical timelines, up-or-out policies, what changes at each level, and how the PSC review process drives promotions.
Last updated: 2026-03-24
Aiming for your next level at Meta?
Get a personalized promotion plan with AI coaching.
Level Overview
| Level | Title | Typical Years | Median TC | Terminal? |
|---|---|---|---|---|
| IC3 | Associate Data Scientist | 1–2 yr | $170K | No |
| IC4 | Data Scientist | 1.5–3.5 yr | $273K | No |
| IC5 | Senior Data Scientist | 2–4+ yr | $443K | Yes |
| IC6 | Staff Data Scientist | 3–5+ yr | $565K | Yes |
Promotion Cycle
Frequency
Annual review with mid-year check-in (moving to 2 full cycles per year under 2026 Checkpoint program)
Decision Maker
hybrid
Manager-driven with calibration committee approval. During the Performance Summary Cycle (PSC), you write a self-review, collect peer feedback from 3-5 nominators, and your manager writes their assessment. The manager then presents your packet at calibration, where it's evaluated against peers at your level. Data scientists are calibrated within the data science org.
Key Details
- •Same PSC process as software engineers — self-review, peer feedback, manager assessment, calibration
- •Review dimensions adapted for DS context: analytical impact, methodology, direction, people
- •Rating scale: Redefines (~3%), Greatly Exceeds (~7%), Exceeds (~35%), Meets All (~45%), Meets Most (~8%), Meets Some (~2%)
- •Two consecutive Meets Most ratings trigger automatic PIP
- •Up-or-out: IC3→IC4 within roughly 24 months, IC4→IC5 within roughly 33 months
- •Manager is the sole advocate in calibration — relationship quality directly affects outcomes
- •Self-advocacy is critical for DS roles — managers may not fully understand the impact of your analytical work
- •2026 Checkpoint program: 2 review cycles per year, bonuses up to 300% for top performers
- •AI usage becoming a core expectation in 2026 reviews — data scientists expected to leverage AI tools
- •External leverage can play a significant role at IC5+ promotions
IC3 — Associate Data Scientist
Entry / New GradEntry point for data scientists at Meta. You run analyses under guidance, build dashboards, write SQL queries, and support product decisions with data. Your manager scopes your projects and reviews your methodology. Up-or-out policy: you have roughly 24 months to reach IC4.
Typical Time at Level
1–2 years (typical: ~1.5 years)
Total Compensation (US)
$145K–$195K (median: $170K)
Source: Levels.fyi
Why Engineers Get Stuck Here
- •Not owning analyses end-to-end — waiting for someone to define the question before you start
- •Staying in SQL/dashboard mode instead of building product sense and proactive insights
- •Weak self-advocacy — doing solid work but not making it visible to your manager and stakeholders
- •Not developing relationships with product and engineering partners
- •Up-or-out clock creates real pressure — roughly 24 months to make IC4
IC4 — Data Scientist
Mid-LevelFirst level of true independence. You own analyses end-to-end, design experiments, build predictive models, and influence product decisions with data. You're expected to identify the right questions, not just answer the ones you're given. Up-or-out policy: roughly 33 months to reach IC5.
Typical Time at Level
1.5–3.5 years (typical: ~2.5 years)
Total Compensation (US)
$230K–$320K (median: $273K)
Source: Levels.fyi
Why Engineers Get Stuck Here
- •Narrow impact — analyses stay within your immediate team and don't scale
- •Not building deep relationships with product and engineering stakeholders
- •Failing to demonstrate IC5-level scope before the promotion decision
- •Calibration opacity — your manager defends your packet in calibration, and results often differ from expectations
- •Meets Most trap — one MM rating triggers a 6-month check-in; two consecutive MM ratings trigger automatic PIP
- •Treating data science as a service function rather than driving product strategy with data
- •Up-or-out pressure — roughly 33 months to reach IC5
IC5 — Senior Data Scientist
SeniorSenior-level scope. You lead complex analytical initiatives, design experimentation frameworks, mentor junior data scientists, and drive strategic decisions across your product area. The up-or-out pressure ends here — IC5 is a terminal level. The jump to IC6 is one of the hardest in data science at Meta.
Typical Time at Level
2–4+ years (typical: ~4 years)
Total Compensation (US)
$370K–$530K (median: $443K)
Source: Levels.fyi
Why Engineers Get Stuck Here
- •Impact scoped to a single team — not generating cross-team analytical influence
- •Not driving strategic data initiatives at the org level
- •Many fail the IC5-to-IC6 promo loop — some consider interviewing directly for IC6 at another company
- •Competing for very limited Staff slots — IC6 promotion is rare and highly contested
- •Not developing the leadership and direction skills that Staff requires
IC6 — Staff Data Scientist
StaffOrganization-wide scope. You design scalable data systems, solve cross-team analytical problems, lead high-impact projects that shape product strategy, and influence without authority. Very selective — getting here from IC5 is one of the hardest promotions at Meta.
Typical Time at Level
3–5+ years (typical: ~5 years)
Total Compensation (US)
$480K–$700K (median: $565K)
Source: Levels.fyi
Why Engineers Get Stuck Here
- •Impact limited to a single product area rather than spanning the org
- •Not shaping analytical strategy or methodology at scale
- •Lacking executive-level visibility for your data-driven initiatives
Additional Context
Meta's data science org uses the same PSC (Performance Summary Cycle) and level system as engineering. Data scientists are leveled IC3–IC7 and face the same up-or-out policies, calibration process, and Meets Most trap as software engineers. The IC5-to-IC6 jump is widely considered one of the hardest promotions in data science — many IC5s plateau here indefinitely. In 2026, Meta is rolling out the Checkpoint program with two review cycles per year and is making AI usage a core performance expectation.
Data sourced from Levels.fyi (compensation, March 2026), Perplexity research synthesis (4 queries), existing Meta company research, Team Blind and Reddit anecdotes. Compensation figures from Levels.fyi. Last verified March 2026.