1. LandOffer
  2. Blog
  3. All articles
  4. Set Boundaries Between the Tech Lead and Engineering Manager Roles

People Management

Set Boundaries Between the Tech Lead and Engineering Manager Roles

A completed fictional TL–EM operating agreement and revised delivery decision retain named technical, people, capacity, product and release authority.

10 min readLandOffer team

Set boundaries between the Tech Lead and Engineering Manager by naming who can decide each recurring question, whose input is needed, and when the decision moves to someone else. Cover technical direction, people management, capacity, delivery commitments and product trade-offs. Titles alone cannot resolve a disagreement, and “shared ownership” is incomplete if neither person knows who must make the call.

This LandOffer.ai guide concerns an existing team, using a completed fictional responsibility map and delivery-conflict record. Public employer guidance was checked October 25, 2026. The examples establish no real authority, reporting line or technical result. They show how a team can document its actual arrangements without assuming every company uses the same TL–EM structure.

Technical-scope and people-capacity planning instruments share a chart table with an escalation bell, illustrating distinct decision responsibilities.

Confirm the actual roles before drawing the boundary

Start with formal reporting lines, position expectations, project assignments and existing approval policies. Is the Tech Lead a temporary project role, a continuing position or also the line manager? Does the Engineering Manager retain technical approval authority? Which decisions belong to product, operations, security or a higher-level leader? Record the answer rather than relying on a familiar title.

GitLab's Tech Lead guidance defines its project role as temporary, distinct from a management position or additional seniority level. It lists possible technical, coordination, planning and risk responsibilities, and explicitly asks the TL and EM to agree who covers the work. That is GitLab's arrangement, not a universal definition of Tech Lead.

Technical authority needs an explicit agreement too. GitLab's Engineering Manager description includes people management and performance feedback alongside technical guidance and final technical decisions when necessary. That illustrates why the specific delegation and escalation route matter. An EM's technical accountability need not require approving every implementation detail, but your team must agree which decisions remain with the manager.

Use a recent ordinary decision to test your understanding: choosing a library, changing a deadline, moving an engineer to another project or supplying performance feedback. If the two leaders describe different approval routes, agree on the route and tell the team before engineers receive conflicting instructions.

Distinguish decision authority from advice and coordination

Separate preparing a recommendation, making the decision and coordinating the work. One person can prepare a recommendation, another can decide, and several can be consulted. The person maintaining the project board may coordinate delivery without having authority to add staff or change business priorities.

For each decision, name the scope and constraint. “Kavi chooses the search component within the approved requirements and architecture constraints” is clearer than “Kavi owns technical decisions.” It leaves a route for an exception affecting another service or requiring a formal review. Similarly, “Robin allocates the team's agreed working capacity” does not authorize Robin to promise another team's engineers.

Clarify how consultation works. If input is required, identify the question and when it is needed. Consultation is not automatically a veto. Conversely, a required approval should not be renamed consultation merely to avoid a disagreement. Existing release, security and personnel processes continue to apply.

Monzo's September 2025 EM expectations describe technical outcomes, delivery, cross-team dependencies and partnership with product management. That company-specific account illustrates overlapping work rather than two isolated territories. A useful boundary makes collaboration easier while retaining a responsible decision maker for each part of a trade-off.

Completed decision map for one engineering team

In this fictional team, Kavi is Tech Lead for a developer-documentation search project. Robin is Engineering Manager and line manager for the engineers. Gita is Product Manager, Vera owns release authorization, and director Ola resolves cross-team priority or capacity exceptions. These assignments are stipulated for the example; another organization may allocate them differently.

The team agrees on this map at the project's start and schedules a review after its first milestone:

Recurring decision Who decides in this agreement Required input and boundary Escalation or handoff
Implementation approach within approved scope Kavi Engineer proposals and existing architecture constraints; no new service commitment assumed Cross-service or architectural exception goes to the designated review owner before commitment
Technical task sequence and acceptance evidence Kavi Implementers estimate their work; Vera supplies release requirements Missing time or people goes to Robin; Kavi does not silently remove required checks
Team capacity and work allocation Robin Kavi explains task needs; engineers identify current commitments Work owned by another team needs its manager's agreement; unresolved conflicts go to Ola
Product scope and priority Gita Kavi describes technical consequences; Robin describes capacity A conflict with an existing organizational commitment goes to Ola
Delivery commitment communicated to stakeholders Robin Gita confirms scope and Kavi supplies the technical forecast Robin returns an unsupported date for a scope, capacity or timing decision rather than promising overtime
Personnel expectations and performance assessment Robin Kavi provides dated observations of work within the company's review process Sensitive concerns use Robin's authorized People or management route; TL feedback is not a private rating system
Technical mentoring and design feedback Kavi for this project Agreed time and learning need; Robin supports development planning A capacity or personnel issue returns to Robin, rather than becoming an indefinite TL obligation
Production release authorization Vera Required technical evidence and operational checks An unmet release condition remains open; neither TL nor EM bypasses Vera's approval

Robin and Kavi also agree that Ola must name any missing formal architecture-review owner before a cross-service option can be approved. The table does not quietly make Ola the approver for every architecture question. A named escalation destination can clarify a missing owner without acquiring all technical decisions itself.

The map includes work outside the pair because many apparent TL–EM conflicts are really unassigned product or release decisions. Kavi cannot settle a customer-priority question by choosing a framework. Robin cannot manufacture technical readiness by approving a staffing plan. Under this fictional agreement, their joint task is to expose the linked questions and send each to its owner.

Keep the team-facing map separate from confidential personnel records. Engineers should understand who gives technical direction, who manages their performance and where conflicting instructions go. They do not need access to colleagues' appraisal details to understand the operating model.

Worked conflict: a technical preference becomes a scope and capacity decision

On May 6 in this fictional project, the team is planning a May 20 milestone. The approved need is finding documentation entries by name and filtering by category. Gita has also requested optional relevance ranking. Kavi recommends a new search service because it could support richer queries later. Robin favors a smaller change within the existing application to avoid creating another service to maintain.

Neither preference is the final decision. Kavi prepares two options using estimates from the implementers. All effort estimates and system conditions here are invented planning assumptions, not measured implementation results or benchmarks.

  • Existing-application option: eleven engineer-days for name search, category filtering and required checks; adding optional ranking would require two further days. No new service is introduced, but richer query behavior remains outside the milestone.
  • New-service option: a preliminary five-week forecast plus a cross-service architecture review and a confirmed maintenance owner. Those approvals and the detailed estimate are not yet available.

Robin checks actual planned availability. One engineer has eight project days; another has three after existing operational coverage. The team therefore has eleven engineer-days allocated. This is an example's agreed plan, not a recommendation to treat every working day as fully interchangeable engineering capacity.

The disagreement is now specific. Kavi can decide an implementation approach within the agreed scope, but cannot commit an unapproved service. Robin can allocate capacity, but cannot drop product scope alone. Gita can remove the optional ranking requirement, but cannot waive Vera's required checks. The May 20 date cannot remain unconditional while those choices are unresolved.

Gita reviews the alternatives and removes optional ranking from this milestone. Kavi selects the existing-application approach for the resulting scope. Robin confirms the eleven-day allocation and communicates a conditional May 20 forecast, with a May 13 checkpoint and release authorization still required. The new-service idea becomes a separate discovery proposal; no production implementation or future staffing is approved merely by recording it.

The conflict is resolved through linked, separately owned decisions rather than a compromise where both leaders remain nominal approvers of everything. Kavi retains technical responsibility for the selected approach. Robin retains the capacity and communication responsibilities. Gita owns the scope reduction, and Vera retains the release gate.

Complete the handoff and test it when the facts change

The team writes the following completed record after the May 6 discussion:

Decision or open item Recorded owner and state Next evidence or action
Milestone scope Gita approves name search and category filtering; ranking is excluded Update the scope statement and stakeholder note on May 6
Implementation approach Kavi selects the existing application within that scope Publish the design rationale and task sequence; normal technical review still applies
Capacity and forecast Robin allocates eleven engineer-days and communicates the conditional date Check remaining work and availability on May 13
Required acceptance and release evidence Kavi coordinates preparation; Vera retains approval No release authorization recorded in the planning decision
New-service proposal Kavi owns a discovery recommendation only Ola identifies the architecture-review route and Gita revisits whether discovery has priority; no work allocation yet

At the fictional May 13 checkpoint, an additional edge case raises the required-check estimate by two engineer-days. Kavi records it instead of trimming the check to preserve the forecast. Robin cannot create those two days without moving another commitment. They ask Gita whether a smaller acceptable scope exists and report the capacity gap to Ola through the agreed route.

The resulting decision is also recorded: Gita keeps both core capabilities; Ola approves moving the internal milestone to May 22, after checking the affected organizational commitment. Robin allocates the first engineer two additional project days, May 21–22, within the now-confirmed schedule and communicates the change. Kavi keeps the required checks in the plan. Vera's eventual release decision remains outstanding.

This revision does not prove the feature is correct or delivered on May 22. It shows a completed planning handoff when the original assumptions changed. The boundary works because a new technical fact reaches the capacity and priority owners without becoming an unauthorized promise or a concealed quality cut.

A shared decision record routes a technical finding to a capacity review while the release-approval space remains open.

Maintain the boundary without building two chains of command

Agree how engineers raise conflicting instructions. Ask them to bring the concrete decision and its effect into the shared project channel or record, so the TL and EM can resolve the conflict before the engineer acts on incompatible directions. The leaders should reconcile their direction and communicate the result together. Employees should not have to negotiate the TL–EM relationship on the team's behalf.

Preserve room for disagreement. Robin can challenge Kavi's reasoning, and Kavi can challenge a forecast that lacks technical support. Neither needs to turn that challenge into a personnel judgment. A TL's observations may inform the manager's assessment through the actual review process; technical disagreement alone does not create a separate authority to rate or discipline colleagues.

Review the map when the project ends, a role changes or recurring decisions keep falling between owners. Cover absences explicitly: a backup can prepare information without automatically inheriting release or personnel authority. If one person fills both roles, label which authority they are exercising and how others can raise an unresolved concern.

Before your next planning conversation, choose one disputed recurring decision and name its owner, required input and escalation condition. Test the agreement against a change in scope or capacity. Keep any missing approver visible, and share the map only with its unresolved decisions clearly labeled. Use LandOffer.ai as the publisher reference for this guide; the immediate task is an operating agreement that matches your actual team and approval rules.

Sources and Further Reading