What is the difference between a director of engineering and a senior engineering manager?

Span and the unit of analysis. Senior-manager: 15–40 people, the unit of analysis is the line-manager and the team. Director: 80–200 people across senior-managers / M3s, the unit of analysis is the sub-org and the multi-quarter strategy. Fournier's chapter 8 ('The Big League') and Larson's 'organizational design' chapter both name the boundary. In practice the boundary depends on company sizing — at smaller companies the title 'Director of Engineering' may apply to a 30-person org, while at FAANG it typically applies to 80+.

How much technical depth does a director need?

Less than M3, but the floor remains 'I can read my orgs' design docs and form judgment.' Charity Majors's charity.wtf framing applies most strongly here: the director who has lost technical credibility loses the ability to defend their org's technical decisions in cross-functional, hire credibly at the senior-IC and management tier, and steer the engineering strategy. By director, technical depth is exercised through judgment about staff/principal IC and senior-manager proposals, not through code review. The dangerous failure mode is the director who hides from technical conversations and becomes a pure people-and-politics function.

Is director the same as VP-Engineering?

Different at most modern-tech-company ladders. Director (D1) is typically the highest-IC-track-equivalent management role; VP-Engineering (D2 / VP-1 in some ladders) is the next rung up, owning multiple director's worth of org. At smaller companies (early-stage startups) the titles are sometimes used interchangeably or with the VP title applied to 60–100 person orgs. At FAANG and large public companies the boundary is clearer. Read the company's leveling rubric or ask the recruiter for the sample scope.

What is the executive-hiring loop at director?

Different shape than IC or line-manager hiring. The behavioral interview round expands (60–90 min, scenario-driven). A panel-style 'walk us through your last re-org' is common. Backchannel references with people who have actually worked with the candidate at the level above are weighted heavily — Horowitz's chapter on hiring executives in The Hard Thing About Hard Things is explicit: 'check references with people who have managed the candidate, not people the candidate selected as references.' The hiring loop is typically 6–10 weeks not 4. Decisions are often made by the VP plus the director plus 1–2 cross-functional peer leaders.

How do directors avoid burnout?

The literature names the same failure mode: directors who do not delegate the M3 / senior-manager job below them and try to be the M3 of every sub-org. Charity Majors writes about this directly on charity.wtf. Larson's An Elegant Puzzle has a chapter on personal effectiveness for senior leaders. The two most-cited mechanics: (1) genuinely delegate the M3 work — hold the M3s accountable for outcomes, not method. (2) Protect 30%+ of the calendar for strategic-thinking time (writing, reading, talking to peers outside the company). The director whose calendar is 100% reactive does not last.

Should a director still do skip-levels with ICs?

Selectively. Larson's 'Skip-level meetings' on lethain.com extends to director: a director typically does skip-skip-levels (1:1s with ICs two levels below) on a 12–18 month rotation, longer for senior ICs and shorter for newer hires. The purpose at director-tier is calibration of the M2/M3s underneath, not direct people-management of the ICs. The frequent failure mode: the director who continues to do every-IC skip-levels at line-manager cadence undermines their senior-managers' authority and burns out.

What is the most common director failure mode?

Misalignment with the VP-Engineering that goes unsurfaced. Horowitz, Lopp, and Larson all name the same pattern: the director who disagrees silently with their VP for two quarters and is then surprised when the VP brings in someone else to lead a critical initiative or asks them to consider a different role. The prevention mechanic: surface disagreement explicitly in 1:1s and in writing. The other major failure modes: not letting go of the M3 job, not delegating people-management to senior-managers, doing too many re-orgs.

Engineering Manager Hub

Director of Engineering / D1 Guide for Tech Companies (2026)

By Blake Crosley · Last verified 2026-04-30

In short

A director of engineering (D1, typically 80–200 people across multiple sub-orgs) is where engineering management transitions from craft to executive function. The D1 partners directly with PM-director, design-director, and engineering-director peers; reports to a VP-Engineering or CTO; and owns multi-quarter strategy, executive-level hiring, and re-org execution. Total comp at FAANG-tier D1 (Meta E8 / Google L8) clusters $900,000–$1.4M with stock vesting per levels.fyi 2026; Anthropic and senior AI-lab engineering directors commonly clear $2M+ on heavy equity. The reading shifts: Larson and Fournier are still core, but Ben Horowitz's 'The Hard Thing About Hard Things' (Harper Business, 2014) and Andy Grove's full HOM become the executive-track references.

Key takeaways

FAANG-tier D1 total comp $900k–$1.4M per levels.fyi 2026; Meta E8 director and Google L8 director sit at this band with stock vesting heavy. Anthropic engineering-director equivalent commonly clears $2M+ on heavy private-company equity. Stripe and Airbnb directors track FAANG closely. (levels.fyi/companies/facebook/salaries/engineering-manager)
The D1 job is executive function: org design, multi-quarter strategy, executive partnership, executive hiring, re-org execution. Larson's An Elegant Puzzle 'organizational design' chapter, Horowitz's 'The Hard Thing About Hard Things', and Grove's High Output Management are the canonical references.
60%+ of the D1 calendar is cross-functional and partnership work. Pragmatic Engineer's coverage of FAANG director schedules and Larson's StaffEng / lethain.com posts both confirm the same pattern: weekly with VP, weekly with peer-PM-director and peer-design-director, weekly staff-of-staff meeting with the senior-managers and M3s underneath, weekly cross-functional with the leadership tier of the partner functions.
Re-org execution is the single hardest D1 craft. Charity Majors's charity.wtf 'sociotechnical path to high-performing teams' and Larson's An Elegant Puzzle organizational-design chapter both treat re-orgs as expensive and consequential. A D1 who does more than one major re-org per 18 months has typically not thought hard enough about the first one.
Executive hiring (senior-managers, GEMs, the occasional director peer) is the most leveraged D1 activity. Larson's 'Hiring senior leaders' (lethain.com) and Horowitz's chapter on hiring executives in The Hard Thing are the canonical references. The hiring loop is heavier on backchannel reference than on interview signal at this tier.
The D1 reports to a VP-Engineering or CTO. The relationship with that boss is the most consequential single relationship in the D1 job. Misalignment with the VP-Engineering produces the canonical D1 failure mode (slow drift, then sudden replacement); strong alignment produces the canonical D1 success pattern (sustained tenure, clear strategy, durable team).
Engineering managers sometimes return to IC at this tier. Charity Majors's 'Engineer/Manager Pendulum' applies even to directors — some directors find the executive function draining and return to staff/principal IC scope. This is not a demotion at most modern-tech-company ladders.

What changes at director: from craft to executive function

The transition from senior-manager / GEM (M2/M3) to director (D1) is the third-largest discontinuity in the EM ladder per the canonical references. Fournier devotes chapter 8 of The Manager's Path to it (titled 'The Big League'). Horowitz's 'The Hard Thing About Hard Things' (Harper Business, 2014) becomes load-bearing reading. Andy Grove's High Output Management — full book, not just the management-leverage chapter — is mandatory. The structural changes:

Span typically 80–200 people across 2–5 senior-managers / GEMs. The unit of analysis is the sub-org, not the team or sub-team. The D1 reasons about which sub-orgs should exist, what their charters are, and how they connect to the company's product strategy.
Most of the calendar is cross-functional and executive partnership. Larson's lethain.com posts on senior-leadership schedules and Pragmatic Engineer's coverage of FAANG director schedules confirm the same pattern: 60%+ of the time is in meetings with peer leaders (PM-director, design-director, engineering-director peers, the VP-Engineering, occasionally the CEO at smaller companies). Direct people-management with the M2/M3s underneath is 20–25%; everything else (strategy writing, hiring, ops review) fills the rest.
Strategy becomes the primary written artifact. Multi-quarter engineering strategy that the M2/M3s underneath execute against. The D1 strategy doc is the parent of the M3 strategy docs underneath. Rumelt's framework still applies but at one level of abstraction higher: the diagnosis is about company strategy and engineering organization fit, not about a single product surface.
Executive-level hiring. The D1 hires senior-managers, M3s, and sometimes director peers. The hiring loop changes shape: a 90-min behavioral interview with the candidate is one signal among many, but backchannel references with people who have actually managed under or alongside the candidate are weighted heavily. Horowitz's chapter on hiring executives in The Hard Thing is the canonical mechanic.
Re-org responsibility. The D1 owns the org structure of their 80–200 people. Mergers of sub-teams, splits of sub-teams, transfers of charter from one sub-org to another. Re-orgs are expensive — Larson's An Elegant Puzzle is explicit — and the D1 who does more than one major re-org every 18 months has likely not thought through the first one carefully enough.

The empirical sign you have made the M3-to-D1 transition: the senior-managers and M3s underneath you run their orgs without daily input; you are spending 60%+ of your time outside engineering with peer leaders; you have written a multi-quarter engineering strategy doc that the company's leadership references; and you have made at least one consequential executive hire that the org credits as a turning point. The empirical sign you have not made the transition: you are still functioning as the M3 of one of your sub-orgs while nominally running the others.

Worked scenario: re-structuring three sub-orgs during a re-org, in 12 months

A 12-month worked scenario — director of engineering owns 4 sub-orgs (140 people total): payments-platform (M3-led, 50 people), payments-product (senior-manager-led, 28 people), risk-and-fraud (M3-led, 35 people), and platform-developer-experience (senior-manager-led, 27 people). The company strategy shifts: the new CEO emphasis is 'reduce time-to-revenue for new use cases', which puts payments-product and risk-and-fraud at the center and pushes payments-platform and developer-experience into supporting roles. The D1 has 12 months to re-structure without losing more than 4 senior leaders. Drawn from Larson's 'Organizational design' chapter (An Elegant Puzzle), Horowitz's chapter on layoffs and reorganizations in The Hard Thing About Hard Things, Pragmatic Engineer's reporting on multiple FAANG re-orgs (newsletter.pragmaticengineer.com), and Charity Majors's writing on org design at charity.wtf.

Months 1–3 (heads-down with VP). The D1 and the VP-Engineering meet weekly to walk through the strategic shift. The D1 maps the current org against the new strategic priorities. The fit is uneven: payments-product is undersized relative to its new importance; payments-platform is correctly sized but its charter overlaps awkwardly with developer-experience. Risk-and-fraud is mostly fine. The VP and D1 align on a re-org thesis: merge developer-experience into payments-platform, expand payments-product by 12 headcount, leave risk-and-fraud alone. The thesis is shared with the CTO and CEO; CEO blesses.
Month 4 (preparation). The D1 has explicit 1:1s with the four leaders affected. The senior-manager who runs developer-experience is the most disrupted: their sub-org is going away and they will report into the payments-platform M3 as a senior-manager. The conversation is direct: this is the strategic call, here is the rationale, here is what your role looks like in the new structure, here is the timeline. The leader has the option to leave with severance per company policy; they choose to stay but ask for clarity on their charter in the new structure. The D1 commits to writing it within 30 days. Horowitz's framing in The Hard Thing About Hard Things: 'tell people the truth, especially when it is hard.'
Month 5 (announcement and the hard week). All-hands. New org structure announced. Within 24 hours every IC who is changing reporting line has a 1:1 with their new manager. The senior-manager whose sub-org is dissolving sends a personal note to each of their reports explaining the change in their own voice. The D1 holds open office hours through the week. Two of the developer-experience ICs decide within the first month that the new structure does not work for them and leave on good terms.
Months 6–9 (execution). Hiring ramp on payments-product. The expanded charter requires three M3-or-senior-manager hires. The D1 spends material time on the leadership pipeline — back-channeling references with their network, conducting the final round of behavioral interviews personally. Two of the three hires land within Q3. The third is a longer search; the D1 holds the line rather than settling.
Months 10–12 (stabilization, retro, second-order effects). Payments-product velocity has materially improved by Q4. Payments-platform's developer-experience absorption has produced one positive effect (cleaner internal API roadmap) and one negative effect (the platform M3 is now stretched thin and is showing signs of burnout). The D1 writes the retrospective with their VP. Top lesson named: under-invested in the M3 stretch — should have hired a deputy under the platform M3 in month 4 rather than month 8. The cost was three quarters of M3 burnout and a near-resignation in month 9.

The lesson Horowitz names directly in The Hard Thing About Hard Things and Larson echoes in An Elegant Puzzle: re-orgs are expensive in trust, in productivity, in human cost. They are sometimes the right answer. They are almost never the right answer twice in 18 months. The D1 who treats re-org as a tool of first resort under-performs; the D1 who treats it as a tool of last resort with named alternatives ruled out first usually nets positive impact.

The VP-Engineering relationship: the most consequential D1 relationship

The single most consequential D1 relationship is with the boss — typically a VP-Engineering, sometimes a CTO. Misalignment with this person is the canonical D1 failure mode (slow drift over 6–12 months, followed by sudden replacement); strong alignment is the canonical D1 success pattern (sustained tenure, clear strategy, durable team). The mechanics, drawn from Horowitz's The Hard Thing About Hard Things, Lopp's Managing Humans (the chapter on managing your manager), Larson's lethain.com posts on managing-up at senior tiers, and Pragmatic Engineer's coverage of director-VP relationships at FAANG-tier:

Cadence. Weekly 1:1, 30–60 minutes. Monthly career and strategic conversation that goes longer. Quarterly written status to the VP that the VP can hand to their boss without modification. The written-status discipline is the second-most-mentioned mechanic across the references (after the 1:1 itself).
Topic balance. The 1:1 should split across: (1) operational status (10–20%), (2) people decisions you are making and want their input on (30%), (3) strategic and cross-functional issues (30%), (4) your own development and career (10–20%), (5) calibration on their thinking and the company's direction (10–20%). The default failure mode is 1:1s that are 80% operational status — that is a misuse of the VP's calendar.
No surprises. Larson's most-cited principle generalizes: the VP should never learn about something serious in your org from someone other than you. Hiring losses, performance issues at the senior-manager+ tier, cross-functional escalations, technical incidents that hit the company's external reputation. Surface-level signal goes in the weekly written status; serious signal goes by Slack or in a same-day 1:1 ping.
Ask explicitly. The D1 who does not ask their VP for explicit feedback on strategic choices and people-management calls is operating without calibration. Lopp's Managing Humans is direct on this: 'your boss is your most reliable signal source about the company's actual priorities, distinct from the stated priorities.'
The pre-mortem on misalignment. If the D1 and VP disagree on a major call, surface the disagreement explicitly in writing and in 1:1. Avoid the most-cited D1 failure mode: the silent disagreement that festers for two quarters and is then revealed at the worst possible time (typically a re-org or perf-cycle moment). Horowitz's chapter on having hard conversations with your boss is the reference.

Compensation: the real bands at director

Total comp at director (D1) FAANG-tier and AI-lab in 2026 (US, per levels.fyi self-reports — director-tier comp has wider variance than IC-tier comp because it depends on company stage, executive equity refresh patterns, and individual negotiation):

Company	Level	Base	Total comp
Meta director-of-engineering	E8	$320k–$420k	$900k–$1.4M
Google director-of-engineering	L8	$320k–$430k	$900k–$1.5M
Stripe director-of-engineering	EM-4 / Director	$330k–$430k	$950k–$1.5M
Airbnb director-of-engineering	L8	$340k–$440k	$1M–$1.6M
Netflix director-of-engineering	Director (entry)	$680k–$850k	$900k–$1.4M (single-band)
Anthropic director-of-engineering	EM-principal	$480k–$600k	$1.8M–$3M+ (heavy equity, peak vesting)
Microsoft Director SDE	69/70	$340k–$430k	$900k–$1.4M
Databricks director-of-engineering	L8	$330k–$430k	$900k–$1.5M

The structural facts at director comp: AI-labs sit furthest above FAANG (Anthropic peak-vesting cycles for engineering directors have produced reported total comp at the top of the engineering-leadership market in 2026); Netflix's single-band model still pushes nominal numbers higher; FAANG director comp is dominated by equity refresh cadence and stock-cycle timing, which makes single-year snapshots noisy. For negotiation: the four-year refresh-grant trajectory matters more than the year-1 number. Pragmatic Engineer's coverage of senior-leadership comp at FAANG is the best public reference for the multi-year framing.

Frequently asked questions

What is the difference between a director of engineering and a senior engineering manager?: Span and the unit of analysis. Senior-manager: 15–40 people, the unit of analysis is the line-manager and the team. Director: 80–200 people across senior-managers / M3s, the unit of analysis is the sub-org and the multi-quarter strategy. Fournier's chapter 8 ('The Big League') and Larson's 'organizational design' chapter both name the boundary. In practice the boundary depends on company sizing — at smaller companies the title 'Director of Engineering' may apply to a 30-person org, while at FAANG it typically applies to 80+.
How much technical depth does a director need?: Less than M3, but the floor remains 'I can read my orgs' design docs and form judgment.' Charity Majors's charity.wtf framing applies most strongly here: the director who has lost technical credibility loses the ability to defend their org's technical decisions in cross-functional, hire credibly at the senior-IC and management tier, and steer the engineering strategy. By director, technical depth is exercised through judgment about staff/principal IC and senior-manager proposals, not through code review. The dangerous failure mode is the director who hides from technical conversations and becomes a pure people-and-politics function.
Is director the same as VP-Engineering?: Different at most modern-tech-company ladders. Director (D1) is typically the highest-IC-track-equivalent management role; VP-Engineering (D2 / VP-1 in some ladders) is the next rung up, owning multiple director's worth of org. At smaller companies (early-stage startups) the titles are sometimes used interchangeably or with the VP title applied to 60–100 person orgs. At FAANG and large public companies the boundary is clearer. Read the company's leveling rubric or ask the recruiter for the sample scope.
What is the executive-hiring loop at director?: Different shape than IC or line-manager hiring. The behavioral interview round expands (60–90 min, scenario-driven). A panel-style 'walk us through your last re-org' is common. Backchannel references with people who have actually worked with the candidate at the level above are weighted heavily — Horowitz's chapter on hiring executives in The Hard Thing About Hard Things is explicit: 'check references with people who have managed the candidate, not people the candidate selected as references.' The hiring loop is typically 6–10 weeks not 4. Decisions are often made by the VP plus the director plus 1–2 cross-functional peer leaders.
How do directors avoid burnout?: The literature names the same failure mode: directors who do not delegate the M3 / senior-manager job below them and try to be the M3 of every sub-org. Charity Majors writes about this directly on charity.wtf. Larson's An Elegant Puzzle has a chapter on personal effectiveness for senior leaders. The two most-cited mechanics: (1) genuinely delegate the M3 work — hold the M3s accountable for outcomes, not method. (2) Protect 30%+ of the calendar for strategic-thinking time (writing, reading, talking to peers outside the company). The director whose calendar is 100% reactive does not last.
Should a director still do skip-levels with ICs?: Selectively. Larson's 'Skip-level meetings' on lethain.com extends to director: a director typically does skip-skip-levels (1:1s with ICs two levels below) on a 12–18 month rotation, longer for senior ICs and shorter for newer hires. The purpose at director-tier is calibration of the M2/M3s underneath, not direct people-management of the ICs. The frequent failure mode: the director who continues to do every-IC skip-levels at line-manager cadence undermines their senior-managers' authority and burns out.
What is the most common director failure mode?: Misalignment with the VP-Engineering that goes unsurfaced. Horowitz, Lopp, and Larson all name the same pattern: the director who disagrees silently with their VP for two quarters and is then surprised when the VP brings in someone else to lead a critical initiative or asks them to consider a different role. The prevention mechanic: surface disagreement explicitly in 1:1s and in writing. The other major failure modes: not letting go of the M3 job, not delegating people-management to senior-managers, doing too many re-orgs.

Sources

About the author. Blake Crosley founded ResumeGeni and writes about engineering management, hiring technology, and ATS optimization. More writing at blakecrosley.com.