Back

Lesson 1/8

6.0

How LinkedIn interviews

1. The shape of a LinkedIn loop

LinkedIn runs a calmer process than Meta but tighter than Microsoft. The loop is a single day, four to five rounds. The pace and decisions come back within a week or two.

A typical senior IC onsite:

Round 1   Coding / problem solving             (45 min)
Round 2   System design                        (60 min)
Round 3   Coding deep dive on a past project   (60 min)
Round 4   Behavioral / culture                 (45 min)
Round 5   Hiring manager + skip-level          (30 min)
A typical LinkedIn onsite

What's distinctive at LinkedIn:

  • The system design round is heavily LinkedIn-flavored. Even when the question is generic ("design a feed"), the interviewer expects you to bring in LinkedIn-specific concerns: connections, jobs, profile relevance.
  • The graph is the secret weapon. LinkedIn engineers think of their product as the Economic Graph — every person, every company, every job, connected. Strong candidates lean into this lens.
  • Calmer interviewer style. LinkedIn historically values collaborative interviews over adversarial ones. Don't mistake this for low standards.

2. What LinkedIn grades you on

LinkedIn's rubric publicly names four things:

SignalWhat it means
Code qualityClean code in the coding rounds; defensible structure in design
Problem-solvingCan you break down a vague problem cleanly?
CommunicationCan you explain technical decisions to non-experts?
Cultural addLinkedIn-specific — "transformation, integrity, collaboration, humor, results, inspiration." Not Amazon-style LP grading, but the words map to real signals.

The cultural-add piece is the underrated one. LinkedIn engineers often work with non-engineering counterparts (product, design, sales). The grade reflects whether you can collaborate across disciplines.

3. What's different here

A few patterns specific to LinkedIn:

  • Graph-first thinking is expected. "Design X" at LinkedIn often translates to "design X on the connection graph." Bringing graph operations into the answer naturally is the senior tell.
  • Multi-tenancy is mild. Unlike Microsoft or Slack, LinkedIn is mostly single-tenant (the global LinkedIn product). But premium tiers (Recruiter, Sales Navigator, Learning) layer on top — design conversations include "how does this differ for a Premium user."
  • Latency budgets are tighter than expected. Profile pages and feed render under 200ms. Interviewers push on the latency budget.
  • The interviewer might be from Recruiter or Sales Navigator. These are LinkedIn's B2B products with different scale and shape than the consumer product. If the interviewer is from these orgs, the questions may shift toward enterprise concerns.

4. The level bar

LinkedIn uses an IC ladder with letter codes. Public level info is sparse:

  • Senior Engineer (~L5 / mid-senior). Most-interviewed level. One deep dive in system design. Single design round sufficient.
  • Staff Engineer (~L6). Two deep dives. Cross-system reasoning expected.
  • Senior Staff (~L7). Roadmap conversations. Migration plans. Organizational implications.
  • Principal (~L8+). Industry-level reasoning. The system design conversations get harder to script — they become real conversations about strategy.

This chapter targets Senior (L5) by default; L6 notes appear where the bar shifts.

5. The LinkedIn-flavored mistakes

6. The seven scenarios + take-home in this chapter

UnitWhat it covers
6.1 LinkedIn feed (job-relevance-aware)Same primitives as Meta's feed; different ranking.
6.2 People You May Know — graph + privacyThe canonical LinkedIn question. Graph BFS with privacy filters.
6.3 Job search and applicant rankingSearch index for jobs; matching with the candidate's profile.
6.4 InMail credits and messagingCredit-budgeted messaging; refill policies.
6.5 Endorsements and skill graphThe skill graph; endorsement aggregation.
6.6 Who viewed your profileCounter + privacy-aware reveal.
6.7 Take-home: connection-strength scoring (code)Build a TS service computing pairwise connection strength.

7. The next page

The LinkedIn feed. Similar to Meta's at the substrate level but with one big difference: ranking is biased toward content from your professional network and toward job-relevant material. We start by separating what's the same and what's different.