Meta Senior Data Engineer Interview (IC5)
Meta (IC5) Senior Data Engineer loop: SQL-heavy with fast-paced coding expectations and a product-sense orientation. Bar at this level: independent technical leadership and cross-team influence. Typical 5-8 years of data engineering experience.
Compensation
$215K–$260K base • $420K–$620K total (IC5)
Loop duration
4.8 hours onsite
Rounds
6 rounds
Location
Menlo Park, NYC, Seattle, London, remote for select teams
Compensation
Meta Senior Data Engineer total comp
Offer-report aggregate, 2019-2026. Level mapped: L5. Typical experience: 7-11 years (median 9).
25th percentile
$286K
Median total comp
$342K
75th percentile
$420K
Median base salary
$200K
Median annual equity
$104K
Median total comp by year
Practice problems
Meta senior data engineer practice set
Meta-tagged practice problems for the senior data engineer level. Click any to start practicing.
7-Day Token Retention
The developer platform team wants a 7-day token retention curve. A token counts as ACTIVE-WITH-TRAFFIC when its status is 'active' (compare it case-insensitively, since the data mixes 'active' and 'Active') and its requests value is greater than 0. A token with no expiration date is treated as still valid. For each issuance date, report two numbers side by side: how many distinct owners had an active-with-traffic token issued that day (active_day0), and how many of those same owners still held an active-with-traffic token that was valid as of 7 days after that date, i.e. issued on or before issued+7 and not expired before issued+7 (active_day7). Order by the issuance date and return only the first 7 dates.
The Dependency Resolver
You're scheduling pipeline tasks from a dependency map where each task name maps to the tasks that must finish before it runs, and every prerequisite is itself a task in the map. Return an order in which every task comes after all of its prerequisites; if the dependencies form a cycle, no such order exists, so raise an exception instead.
Content Engagement Data Model
We run a large social content platform where creators publish posts (text, image, video) and people engage through views, reactions, comments, and shares. Each view carries how long the person watched and each reaction carries its kind (like, love, angry), while a comment can reply to another comment so reply threads nest. A share turns an existing post into a brand-new post of its own, so the model has to trace every reshare back to both the post it came from and the person who reshared it. Design a schema that powers virality, creator-performance, and feed-ranking dashboards, knowing that raw engagement runs to billions of events a day, far too much to scan on every dashboard load.
Cheapest CDN Route
The CDN team is sizing the object cache at the SFO edge and wants the heaviest thing it has ever delivered successfully from there. Among status-200 responses served from SFO, return the largest byte payload as a single value.
Walk into Meta knowing the system design pattern they'll test.
Rolling 7-day active users
Count distinct users active in the trailing 7 days for each date. Product analytics staple.
Two Million Boxes by Monday Morning
Shipped, maybe. Delivered, debatable.
Pulled from debriefs where system design separated levels.
The loop
How the interview actually runs
01Recruiter screen
30 minNon-technical. The recruiter confirms level, product area (Ads, Integrity, Instagram, Reality Labs), and motivations. How you describe past work signals IC3/IC4/IC5.
- →Quantify everything: row counts, daily event volumes, TB processed
- →Research the specific team. Meta has dozens of DE teams with different tech stacks
- →Ask whether the loop includes a Python round; some teams do, some don't
02Technical phone screen
45 minLive SQL coding, 1-2 problems, in a shared doc with no syntax highlighting. Problems emphasize window functions, multi-step logic, and event-stream schemas.
- →Think out loud from the start, silence worries the interviewer
- →Expect window functions: ROW_NUMBER, LAG, LEAD, running totals
- →Ask clarifying questions: NULL handling, duplicates, timezone of timestamps
03Onsite: SQL deep-dive
45 min2-3 SQL problems with increasing complexity. The last often adds an optimization discussion: 'Your solution works, now make it efficient on 500B rows.'
- →Practice writing SQL without autocomplete. Meta uses a shared doc
- →When discussing optimization, mention partition pruning, predicate pushdown
- →Use CTEs to break complex queries into readable steps
04Onsite: Python / data manipulation
45 minPractical data work, not LeetCode. Parse JSON logs, transform nested structures, write a data validation function, build a small ETL step.
- →Practice file I/O, dictionary manipulation, list comprehensions
- →Write helper functions instead of one monolithic block
- →Handle edge cases explicitly, empty inputs, missing keys, malformed data
05System design (pipeline architecture)
60 minDesign a production pipeline end-to-end: ingestion, transformation, storage, consumers, SLAs, failure modes, backfill strategy, and cost trade-offs. At senior level, you drive the conversation without prompting. Expect follow-ups about scale, cross-team coordination, and operational load.
- →Anchor on the SLA and data shape before diagramming
- →Discuss idempotency, partitioning, and backfill explicitly
- →Estimate cost: 'This pipeline will cost roughly $X/month at this volume'
Level bar
What Meta expects at IC5 Senior Data Engineer
Independent technical leadership
Senior DEs drive pipeline designs without engineering manager involvement. Interviewers probe whether you can decompose ambiguous requirements, make architecture trade-offs, and defend your choices under scrutiny.
Cross-team coordination
Senior scope regularly spans multiple teams. Expect scenarios about a downstream team missing an SLA because of a change you made, or negotiating a schema migration with the team that owns the upstream source.
Production operational rigor
Fluent in on-call, alerting, data quality checks, and incident response. Dive-deep stories at this level should include correlating a metric drop to a specific commit or a timezone bug or a subtle ordering issue, not 'I looked at the logs.'
Meta-specific emphasis
Meta's loop is characterized by: SQL-heavy with fast-paced coding expectations and a product-sense orientation. Calibrate your preparation to that, generic FAANG prep will not close the gap on company-specific expectations.
Behavioral
How Meta frames behavioral rounds
Move Fast
Meta's culture rewards shipping and iterating. Stories about shipping a V1, measuring, and iterating land harder than stories about getting a design perfect before launch.
Focus on Long-Term Impact
Paired with Move Fast. Meta wants DEs who ship fast without creating 3-year tech debt. Balance matters.
Build Awesome Things
Meta wants people who care deeply about craft. Your ETL pipeline is not just a job, it is a thing you built.
Live in the Future
Senior and above: betting on the technology curve. Candidates who talk about where data infrastructure is going in 3 years land strongly.
Prep timeline
Week-by-week preparation plan
Foundations and gap analysis
- ·Do 10 medium SQL problems. Note which patterns feel slow
- ·Write out 2-3 behavioral stories per value, Meta weights this round heavily
- ·Read Meta's public engineering blog for recent architecture patterns
- ·Review your prior production work, pick 3-5 projects you can discuss in depth
SQL and coding fluency
- ·Practice window functions until DENSE_RANK, ROW_NUMBER, LAG, LEAD are reflex
- ·Do 20+ Meta-style problems in their domain
- ·Time yourself: 25 min per medium, 35 min per hard
- ·Record yourself narrating your approach aloud. Interviewers weigh how you explain it, not only what you write
Pipeline system design
- ·Design 5 pipelines on paper: daily aggregation, clickstream, CDC, ML feature store, real-time alerting
- ·For each, write SLA, partition strategy, backfill plan, and cost estimate
- ·Practice with a friend, senior-level system design is 50% driving the conversation
- ·Review Meta's open-source and engineering blog for in-house patterns
Behavioral polish and mock loops
- ·Rehearse every story out loud. Cut to 2-3 minutes each
- ·Run 2 full mock loops with a senior DE or coach
- ·Identify your 3 weakest behavioral areas and draft additional stories
- ·Review recent Meta news or earnings call for fresh talking points
Taper and logistics
- ·No new content. Review your notes only
- ·Sleep. Mental energy matters more than one more practice problem
- ·Confirm logistics: laptop charged, shared-doc tool tested, snack and water nearby
- ·Remember: the loop is rooting for you to raise the bar, not to fail
See also
Related pages on Meta's loop
FAQ
Common questions
- What level is Senior Data Engineer at Meta?
- On Meta's ladder, Senior Data Engineer sits at IC5. Expectations center on independent technical leadership and cross-team influence.
- How much does a Meta Senior Data Engineer make?
- Across 322 offer samples from 2019-2026, Meta Senior Data Engineer total compensation lands at $286K (P25), $342K (median), and $420K (P75), median base $200K and median annual equity $104K. Typical experience range: 7-11 years..
- How is the Senior Data Engineer loop different from other levels at Meta?
- Round structure is shared across levels; what changes is what each round tests. For Senior Data Engineer the emphasis is independent technical leadership and cross-team influence, with particular attention to independent system design and cross-team influence.
- How long should I prepare for the Meta Senior Data Engineer interview?
- 8-10 weeks of focused prep is typical for candidates already working as a DE. Less than 4 weeks is tight; the behavioral story bank usually takes longer than candidates expect.
- Does Meta interview data engineers differently than software engineers?
- Yes. DE loops at Meta weight SQL heavier, include pipeline/system-design rounds tuned to data workloads, and probe for production data experience (ingestion patterns, data quality, backfill) that generalist SWE loops skip.
50+ guides covering every round, company, and role