Meta Site Reliability Engineer Interview Guide

The short answer: Meta does not post a Site Reliability Engineer title. Its SRE-type role is the Production Engineer, which Meta's postings describe as a specialized software engineer who keeps services reliable and scalable and shares on-call. Meta says its process has a recruiter conversation, an initial screening and a full loop, and typically takes about 2 to 3 months. Candidate reports describe a screen with separate coding and troubleshooting interviews, then a loop of coding, system design, Linux and systems internals, and behavioral rounds. Meta says no outside AI tools are authorized and only the tools inside the interview environment may be used. Last reviewed October 2026.

Meta site reliability engineer interview: quick answers

The Meta site reliability engineer role

What Meta site reliability engineers work on, the skills its job postings ask for, and what a week can look like. Meta posts the role as Production Engineer and Production Engineering (Meta posts no Site Reliability Engineer title).

Meta site reliability engineer levels and pay

Your level sets the bar you are interviewed against and the pay band of the offer.

Meta's PE postings do not name levels. They are split by experience: Production Engineering asks for 2+ years of *nix, networking and coding experience, and Production Engineer asks for 7+ years plus mentoring. One 2025 candidate interviewed for a role labeled E4.

Pay in Meta's US job postings

Meta's two US Production Engineer postings each list one base salary range for the United States. Both mention bonus and equity on top of base.

The postings name no levels, so the rows use the posting titles and their experience bars. The intern posting, which lists monthly pay, is left out. Ranges seen October 29, 2026.

Meta site reliability engineer interview process

Meta publishes one hiring process for all roles and no Production Engineer interview guide, so the steps follow its hiring process page. The content of each PE round comes from candidate reports: one detailed 2025 report and a 2024 thread.

Based on Meta's own pages and candidate reports.

Meta says the typical process takes about 2 to 3 months, and one 2025 candidate reported about three months from December to March. Only one detailed full-loop report was readable, so treat the round list as a likely shape and confirm it with your recruiter. Last reviewed October 2026.

What is different at Meta

• The title is Production Engineer. Meta posts no Site Reliability Engineer title, and its PE postings describe specialized software engineers who write and review code, own back-end services and debug live production. Network Production Engineer is a separate network track, and Production Systems Engineer roles are data center and hardware fleet jobs.
• PEs sit inside product and core infrastructure teams. Meta names Instagram, WhatsApp, Oculus and Videos, and core services such as Storage, Cache and Networking, including some of the world's largest MySQL deployments. Production engineering is listed under Meta's Infrastructure organization.
• On-call is shared. Postings say PEs share an on-call rotation with their team and act as an escalation contact for service incidents. They also write capacity plans and documentation alongside code.
• The screen includes troubleshooting. Two candidate reports describe a dedicated troubleshooting interview next to coding in the screen, and the coding questions in reports lean toward practical file parsing.
• AI shows up in the job and in the rules. PE postings list prompt and context engineering, agent orchestration and using AI tools to ship code as preferred skills. For interviews, Meta says no outside AI tools are authorized and only the tools inside the interview environment may be used.

Meta site reliability engineer interview questions

Questions from our bank, then practice prompts we wrote for this guide around Meta's work.

1from typing import List, Tuple
2
3def disks_filling(csv_text: str, hours: float) -> List[Tuple[str, str, float]]:
4    """Return (host, mount, hours_left) for mounts projected to fill within `hours`."""
5    pass

1import csv
2import io
3from typing import List, Tuple
4
5def disks_filling(csv_text: str, hours: float) -> List[Tuple[str, str, float]]:
6    samples = {}  # (host, mount) -> list of (timestamp, used_gb, size_gb)
7    for row in csv.DictReader(io.StringIO(csv_text)):
8        try:
9            host, mount = row['host'].strip(), row['mount'].strip()
10            ts, used, size = int(row['timestamp']), float(row['used_gb']), float(row['size_gb'])
11        except (KeyError, TypeError, ValueError, AttributeError):
12            continue  # missing or non-numeric field
13        if not host or not mount or size <= 0 or used < 0:
14            continue
15        samples.setdefault((host, mount), []).append((ts, used, size))
16
17    result = []
18    for (host, mount), rows in samples.items():
19        rows.sort()
20        t0, u0, _ = rows[0]
21        t1, u1, size = rows[-1]
22        if u1 >= size:
23            result.append((host, mount, 0.0))
24            continue
25        if t1 == t0:
26            continue  # one sample: no growth rate
27        rate = (u1 - u0) / (t1 - t0) * 3600  # GB per hour
28        if rate <= 0:
29            continue
30        left = (size - u1) / rate
31        if left <= hours:
32            result.append((host, mount, round(left, 1)))
33    return sorted(result, key=lambda r: (r[2], r[0], r[1]))
34
35# Complexity: O(n log n) time for sorting samples per mount, O(n) space.

How to prepare

A plan for the Meta loop, most important first:

1. Practice Linux troubleshooting on a real machine: fill a disk, exhaust inodes, leak a file descriptor, then find each with standard tools.
2. Practice practical coding: parse CSV and log files in Python, handle bad rows, and extend the solution when the interviewer adds requirements.
3. Review systems internals: memory management and the OOM killer, processes and signals, file systems and inodes, permissions and TCP basics.
4. Prepare PE-style designs with numbers: capacity plans, failover between data centers, caching in front of MySQL and safe rollouts.
5. Write behavioral stories on collaboration, conflict, technical leadership and an incident where you were the escalation contact.
6. Ask your recruiter which rounds your loop has and whether it uses the built-in AI assistant, and use the practice session in your Career Profile.

Practice path

Courses that cover the rounds above:

• Site Reliability Engineering Interview
• Python Basics
• System Design
• Data Structures and Algorithms