Engineering Roles

Purpose: Define the core roles across Engineering (and critical adjacent partners) with crisp missions, decision rights, operating loops, KPIs, and “definition of done.” Ends with a RACI over the most common responsibilities.

Default assumptions: single paved road; ringed deploys; SLOs before GA; one owner per service; WIP ≤ 2 bets/squad.

1) CTO — Chief Technology Officer

North Star: Turn strategy into shipped, reliable outcomes.

Owns (A): Org topology; quality bars policy (flags/telemetry/rollback/SLOs); architecture principles; reliability posture; talent system; board/executive tech narrative.

Weekly loops: Outcome review; reliability huddle; platform council; hiring loop; 1:1s with leads.

KPIs: Lead time p50 < 24h · CFR < 15% · MTTR < 1h · SLO attainment ≥ 99% · Platform adoption ≥ 80%.

Definition of Done: 100% services owned with SLOs & runbooks; policy gates enforced in CI; two months of green DORA+SLO; cadence running (weekly/monthly/quarterly).

2) HoE — Head of Engineering / VP Eng

North Star: Predictable, high‑quality delivery.

Owns (A): Delivery ops (WIP caps, change calendar), gate enforcement, service catalog hygiene, incident program, hiring/onboarding execution, dependency management.

Weekly loops: WIP & deps sweep · team outcome reviews · reliability huddle · platform check · hiring/people.

KPIs: Lead time < 24h · deploy ≥ daily · CFR < 15% · MTTR < 1h · wait time ≤ 2d · adoption ≥ 80%.

Definition of Done: Ops scoreboard live; gates enforced in CI; pages ≤2/team/wk; adoption trending to 80%.

3) HoP — Head of Platform

North Star: A paved road that makes the right way the easy way.

Owns (A): Templates/scaffolds; CI/CD; preview envs; observability defaults; platform SLOs; deprecations; supply chain (SBOM/scanning); internal customer success (SLA/docs/office hours).

Weekly loops: SLO dashboard review · backlog triage · office hours · migrations check.

KPIs: Build p50 < 10m (p95 < 20m) · Preview p95 < 5m · Uptime ≥ 99.9% · Flake < 2% · Adoption ≥ 80% · Dev NPS ≥ +40.

Definition of Done: Charter+SLOs published; SLOs green; ≥60% adoption and rising; exception memos tracked; deprecation wave completed with guides.

4) Stream TL — Tech Lead (value‑stream squad)

North Star: Ship outcomes quickly and safely for one stream.

Owns (A): End‑to‑end of stream services (SLOs, runbooks, on‑call); testing strategy; ringed deploys; within‑boundary architecture; security/privacy for changes.

Weekly loops: Outcome review (with PM) · WIP≤2 & dependency pass · reliability huddle · post‑deploy verification · tech‑debt/perf slice.

KPIs: Lead time < 24h · deploy ≥ daily · CFR < 15% · MTTR < 1h · SLOs green · wait time ≤ 2d.

Definition of Done: One owner per service; contract tests for external interfaces; rollback < 5m; preview p95 < 5m; pager quiet (≤2/wk).

5) PM — Product Manager (stream‑aligned)

North Star: Move the NSM and KRs with the fewest, safest changes.

Owns (A): Problem briefs & success metrics; within‑stream portfolio (WIP ≤ 2); experiment plans (MDE/power/guardrails); readiness gates; release notes & enablement; outcome reviews.

Weekly loops: Outcome review · backlog/WIP check · customer time · evidence & decision notes · GTM sync.

KPIs: KR delta/week on target · activation/adoption thresholds · idea→decision ≤ 2w · CFR ≤ +2pp vs baseline post‑release.

Definition of Done: Problem briefs, experiment plans, decision notes, KR tree, outcome dashboard; one Pilot→Beta→GA completed with guardrails intact.

6) SRE Lead / SRE

North Star: Keep systems available, fast, and safe—predictably.

Owns (A): SLOs/SLIs/alerts; error‑budget policy; incident response (on‑call, RCAs, drills); observability; resilience controls; DR/backups.

Weekly loops: SLO & alert review · incident huddle · change‑safety sync · on‑call health.

KPIs: SLO periods green ≥ 99% · MTTR < 1h · pages ≤ 2/team/wk · alert precision > 90% · backup restore 100%.

Definition of Done: SLOs live for tier‑1/2; burn alerts wired; policies gating rings; DR drill evidence; RCAs closed with actions merged.

7) Data Lead / Analytics Engineer

North Star: Trusted, self‑serve data and decision‑quality experiments.

Owns (A): Metric definitions & semantic layer; event taxonomy & data contracts; data reliability (freshness SLAs, tests, lineage); self‑serve BI & governance; experiment standards.

Weekly loops: Freshness & tests review · schema/contract triage · experiment clinic · office hours.

KPIs: Metric spec coverage 100% · freshness ≥ 99% · tests ≥ 95% pass · data MTTR < 1h · semantic‑layer usage ≥ 80%.

Definition of Done: NSM/KR specs live; SLAs green; certified dashboards; experiment policy enforced (MDE/power/guardrails).

8) Staff / Principal Engineer (senior IC)

North Star: Raise the technical bar and unlock multiple teams.

Owns (A): System design in a domain; one‑way door choices (ADR/RFC); reference implementations & golden paths; performance budgets; deprecations & migrations; design review program.

Weekly loops: Design reviews · pairing/implementation · reliability & perf triage · mentoring · short docs/ADRs.

KPIs: Wait time ≤ 2d · CFR < 15% · MTTR < 1h · hot‑path p95 on budget · golden‑path adoption ≥ 80%.

Definition of Done: Two cross‑team wins shipped; contracts in CI; perf budgets gated; deprecation plan executed; review SLA ≤ 48h.

9) Software Engineers — Stream & Platform

North Star: Ship small changes that move KRs, safely.

Core responsibilities (R):

• Build on the paved road; keep PRs small; add tests at the right layer.
• Operate what you ship: SLOs, alerts, runbooks, on‑call rotation.
• Instrument features (events, metrics) and annotate deploys.
• Use contracts and versioning for inter‑team interfaces; keep provider/consumer tests green.
• Write helpful notes (docs/checklists) and follow post‑deploy verification.

Weekly habits: 5–10 PRs merged; post‑deploy smoke; fix one paper‑cut; pair once; update one runbook/checklist.

Quality bars: No secrets in code; rollback < 5m; preview p95 < 5m; zero new flaky tests; changes observable by default.

RACI — Core Responsibilities by Role

Roles: CTO · HoE · HoP · TL · PM · SRE · Data · Staff/Princ · Eng

RACI key: A = Accountable (single owner); R = Responsible (executes); C = Consulted (two‑way input); I = Informed (kept in the loop).

How to use this

• Link this doc from your Service Catalog and Team APIs.
• For any responsibility with unclear A/R, resolve in 1:1s or at the Platform/Outcome council this week.
• Keep the RACI to one A per row; review quarterly or after topology changes.

Glossary

Alphabetical; concise definitions tuned to the context of the guide.

A

• ADR — Architecture Decision Record: Short note that records a key technical decision, its context, and consequences.
• API — Application Programming Interface: A contract for how software components communicate. (Mentioned implicitly via “interfaces/contracts”.)

B

• BI — Business Intelligence: Tools/dashboards used for analysis and decision‑making.

C

• CD — Continuous Delivery/Deployment: Automating release to staging/production safely and frequently.
• CFR — Change Failure Rate: % of deployments that cause incidents/rollbacks.
• CI — Continuous Integration: Automated build/test on every change.
• CoD — Cost of Delay: Economic impact of delaying a feature/outcome.
• CTO — Chief Technology Officer: Exec owning technology strategy and outcomes.

D

• Data MTTR: Mean time to recover from data incidents. (Data reliability context.)
• Dev NPS — Developer Net Promoter Score: Satisfaction measure from internal developers.
• DORA — DevOps Research & Assessment (metrics): Lead time, deploy freq, change failure rate, MTTR.
• DR — Disaster Recovery: How systems are restored after major failure (see RTO/RPO).
• DX / DevEx — Developer Experience: How easy and pleasant it is for engineers to build/ship. (Implied by Platform/DevEx targets.)

F

• FinOps — Cloud Financial Operations: Practices to manage cloud/infrastructure spend.

G

• GA — General Availability: Broadly available, production‑ready release.
• GTM — Go‑to‑Market: Coordination with marketing/sales/support for launches.

H

• HoE — Head of Engineering (VP Eng): Runs delivery operations and engineering execution.
• HoP — Head of Platform: Owns the paved road (templates, CI/CD, previews, observability).

I

• IC — Individual Contributor: Non‑manager engineer role.
• IR — Incident Response: Process/roles to handle outages/security events.

K

• KPI — Key Performance Indicator: Metric tracking health/performance of an area.
• KR — Key Result: Outcome target aligned to the NSM.

M

• MDE — Minimum Detectable Effect: Smallest effect size an experiment aims to reliably detect.
• MTTR — Mean Time To Restore (or Recover): Average time to mitigate/resolve incidents.

N

• NPS — Net Promoter Score: Satisfaction metric (here, among internal developers for platform).
• NSM — North Star Metric: Single most important metric guiding the product/business.

O

• Ops — Operations: Practices and routines for running systems and teams.

P

• p50 / p95: 50th/95th percentile of a metric (e.g., build time, latency).
• PII — Personally Identifiable Information: Data that can identify an individual.
• PM — Product Manager: Owns problem framing, outcomes, portfolio/WIP inside a stream.
• PR — Pull Request: Proposed code change reviewed before merge.

R

• RACI — Responsible, Accountable, Consulted, Informed: Model clarifying who does what.
  • R — Responsible: Executes the work.
  • A — Accountable: Single owner answerable for the result/decision.
  • C — Consulted: Gives input before decisions.
  • I — Informed: Kept up to date after decisions.
• RFC — Request for Comments: Design document used to discuss/decide significant changes.
• RICE — Reach, Impact, Confidence, Effort: Framework to prioritize bets/initiatives.
• ROI — Return on Investment: Value delivered relative to cost.
• RPO — Recovery Point Objective: Maximum acceptable data loss window in DR.
• RTO — Recovery Time Objective: Target maximum downtime in DR.

S

• SaaS — Software as a Service: Cloud‑hosted software delivered via subscription.
• SBOM — Software Bill of Materials: Inventory of components/dependencies in a build.
• SLA — Service Level Agreement: Promised service targets (often contractual).
• SLI — Service Level Indicator: Measured signal of service health (e.g., latency).
• SLO — Service Level Objective: Target for an SLI (e.g., 99.9% availability).
• SRE — Site Reliability Engineering / Engineer: Discipline/role focused on reliability.
• SLSA — Supply‑chain Levels for Software Artifacts: Framework to harden software supply chains. (Referenced via supply‑chain posture.)
• SAST/DAST: Static/Dynamic Application Security Testing. (Security scanning in CI/CD.)

T

• TL — Tech Lead: Leads a stream‑aligned squad across design/build/test/release/operate.

V

• VP — Vice President. (As in VP Engineering.)

W

• WIP — Work in Progress: Concurrent bets/initiatives; capped to protect flow.

Y

• YAGNI — “You Aren’t Gonna Need It”: Principle to avoid premature abstraction/over‑engineering.

Notes

• Percentiles (p50/p95) and DORA are used as SLIs/KPIs to track flow, quality, and reliability.
• RACI rows in the guide intentionally have one “A” per responsibility to avoid decision ambiguity.