The Leveling Problem
Titles are meaningless without calibration. If "Senior Engineer" means different things on the Mobile team versus the Infrastructure team, you have created a system that rewards gaming over merit. Engineers will seek out the team with the lowest bar for promotion, titles will inflate, and compensation bands will become impossible to manage fairly.
More importantly, uncalibrated levels make hiring impossible. If the hiring bar for "Senior" varies by team, you cannot train interviewers, you cannot write consistent rubrics, and you cannot make fair comparisons between candidates.
Calibration Sessions
The solution is regular calibration sessions — typically quarterly — where managers from different teams align on a shared definition of each level. The format:
- Anonymized examples: Managers bring 2–3 anonymized descriptions of work done by people at each level on their team.
- Cross-team comparison: The group discusses whether the work described matches the level claimed.
- Friction as signal: Disagreement is the point. When Manager A says "This is clearly Senior work" and Manager B says "This looks Mid-level to me," the conversation that follows produces the shared understanding that calibration is designed to create.
Example calibration exchange:
Manager A: "Jane shipped our new CI/CD pipeline. I think that's Senior-level work."
Manager B: "What was the scope? Was it greenfield or migrating an existing system?"
Manager A: "Greenfield, but it was a standard Jenkins setup."
Manager B: "On my team, a Senior has to make architectural trade-off decisions under ambiguity — choosing between tools, justifying the decision to stakeholders, handling edge cases. A standard Jenkins setup is solid Mid-level work."
Through this friction, the organization builds a shared vocabulary. "Senior" stops meaning "has been here three years" and starts meaning "demonstrates these specific capabilities."
Calibration and Hiring
Calibration sessions directly improve hiring quality because they force the organization to articulate what it is actually looking for. The output of a calibration session — a concrete description of what each level looks like in practice — becomes the input to the interview rubric. This closes the loop: the bar you hire to is the same bar you promote to, which is the same bar you calibrate against.
Engineering managers are frequently asked: "How do you ensure consistency in how you evaluate candidates across your team?" The answer interviewers want is not "I trust my interviewers." It is a description of a system: structured loops with assigned competencies, pre-defined rubrics, blind feedback submission, and regular calibration sessions that align the team on what each level actually requires.