The Shift
What happened to velocity? Agents multiplied individual throughput by three to ten times, and the metric that defined engineering management for a decade turned into noise. Story points lost meaning when an agent could close a backlog’s worth of tickets overnight. The classic EM, measured by sprint burndown, standup hygiene, capacity plans, and a steady stream of “unblocked” tickets, was optimizing for a constraint that evaporated.
You’re not managing eight engineers anymore. You’re managing eight engineers and their agents, each pair producing output that legacy metrics can’t interpret. The EM who keeps optimizing for volume is managing the wrong era.
The job moves to what HELM encodes across the organization. Shift 2 (Leadership): stop directing execution, start defining constraints so teams and agents operate inside clear boundaries. Shift 6 (People systems): measure impact, not output volume. Principle 6 (Team-Wide Adoption Over Individual Mastery): the team’s effective capability is set by its least-adopted critical-path member, not its star performer. The engineering manager executes those shifts daily: building capability, balancing adoption, and insisting the team measures what actually matters.
What the Traditional Job Description Looked Like
Every EM posting read the same way. Five-plus years managing engineering teams. Agile and Scrum fluency. Velocity tracking, capacity planning. Strong stakeholder communication. Often a story about scaling headcount from X to Y. Interviews rewarded backlog grooming, sprint planning, and escalation paths. The emphasis was process management: keeping the train on the tracks and the metrics green.
The Transformed Role
Core Mission
Build and lead an AI-augmented engineering team where every member operates effectively with agents, the team measures impact over output, and adoption levels are balanced across the pod.
Key Responsibilities
- Assess and advance the team's maturity level using the HELM Maturity Model (Levels 1–5)
- Ensure adoption equity: treat the team's effective level as the least-adopted member in a critical-path role, and close that gap deliberately
- Define and enforce the team's operating rhythm around the Plan-Execute-Verify-Ship-Learn loop
- Restructure workflows from horizontal silos to vertical, outcome-oriented cross-functional pods
- Establish and track KPIs that measure product impact, not output volume—lead time, change failure rate, escaped defects—not story points or pull requests merged
- Coach engineers on judgment, review quality, and context engineering, not only on narrow technical skills
- Own execution of the adoption roadmap (Phases 1–4) for the team, with clear exit criteria per phase
- Manage the human side of AI adoption: resistance, anxiety, identity shifts, and the honest question—what is my job now?
Required Competencies
- Adoption management — Assessing where each person sits on the maturity model and building individual development plans that raise the team floor, not only the ceiling.
- Outcome-based measurement — Designing KPI frameworks that reward impact over activity, aligned with the HELM KPI Dashboard.
- Organizational design — Moving teams from functional silos to pods organized around outcomes and customer value.
- Change leadership — Holding space for emotional and identity challenges as roles transform, without letting avoidance stall the whole organization.
- Coaching for judgment — Developing engineers who can evaluate and review at volume with consistent standards.
- Agentic workflow fluency — Knowing the Plan-Execute-Verify-Ship-Learn loop well enough to see where it breaks in practice and fix the system, not blame individuals.
What We No Longer Screen For
- Ability to "run a good standup" or optimize Scrum ceremonies as the core proof of management skill
- Deep expertise in a specific project-management toolchain (e.g., Jira wizardry) as a proxy for leadership quality
- Velocity tracking and capacity planning anchored in story points
- Team growth framed primarily as headcount scaling
- Performance management driven mainly by code output or merge volume
How We Interview
- Adoption equity scenario — "Your team has one Level 4 engineer and two Level 1 engineers. The Level 4 engineer is frustrated that their pull requests wait for review. What do you do?" Tests whether you protect stars or lift the floor and rebalance how work flows.
- Measurement design — "Design a KPI framework for a team adopting agents. What do you measure, what do you stop measuring, and why?"
- Change management — "An experienced senior engineer says 'AI is going to replace us all' and resists adoption. How do you handle it?"
- Team design — "You are restructuring from a frontend team and a backend team into vertical pods. Walk through your approach—stakeholders, sequencing, and risk."
- Maturity assessment — Given a concise description of current practices, place the team on the maturity model and propose a 90-day advancement plan with milestones and exit criteria.
Day in the Life
A senior engineer pings you: "I approved six agent PRs yesterday and I'm not sure three of them were good enough." That's your signal. The 1:1 becomes a coaching session on review heuristics, sampling strategies, and when to escalate. You're building skill, not assigning more tickets.
Later, you check team-level signals. Lead time, change failure rate, escaped defects, adoption KPIs. Not a burndown chart. The question that matters: is the system getting safer and faster in ways users actually feel, or is adoption concentrating in two people while everyone else waits?
Pod sync is brief. Where is the loop stuck, who needs pairing, what rule or template should change. No point totals. Then you spend time on the 90-day adoption roadmap: which phase the team is in, what evidence would satisfy exit criteria, and what you'll do this week to move the lowest-adopted critical-path person forward.
Connection to HELM
The engineering manager is where HELM's Six Organizational Shifts meet the calendar. You translate the Leadership Guide into weekly choices about constraints, measurement, structure, and people systems. The Maturity Model (Leadership and Practitioners guides) gives you a shared language for where the team is and what "better" means next. You own execution of the Adoption Roadmap: phases, gates, and honest assessment of whether the team is ready to deepen autonomy. The KPI Dashboard keeps the team and leadership aligned on impact instead of vanity metrics. Principle 6: Team-Wide Adoption Over Individual Mastery is the mandate in one line: the pod wins when everyone in the critical path can ship with agents at the required bar, not when one hero carries the load.