1334 questions
Could not save this preference. Your choice applies only to this page.
Stage profile · Individual contributor
Owns a feature or component end to end, delivers reliably, follows operational feedback, and supports peers when relevant to the role.
Read these as observable expectations across repeated work, not as a checklist for a single project. First distinguish whether the role expects the work, whether the person had a chance to demonstrate it, and whether the evidence supports the claim. Lack of opportunity or evidence is not itself a negative judgment. AI use is judged by verified outcomes, sound boundaries, and the quality of decisions; choosing not to use AI can be sound.
These lenses and the A–H capability map answer different questions. One experience may provide evidence across lenses, but do not count the same result multiple times.
Understand the people served and intended outcomes, make evidence-based trade-offs, and verify contribution. For platform, infrastructure, reliability, and safety work, cost, efficiency, risk reduction, and internal-user outcomes count; direct revenue ownership is not universal.
Apply relevant knowledge to diagnose, design, implement or review, validate quality, and make evolution decisions. The professional leaves required depend on the role.
Help others work effectively through communication, coordination, review, shared methods, coaching, or organizational mechanisms. IC influence is not measured by direct reports; managers also own formal people responsibilities.
Observe questioning, reasoning, learning transfer, judgment updates, relationship handling, and recovery in work. This is not a fixed personality or potential verdict; do not infer it from extroversion, fluent speech, or willingness to overwork.
Use this structure to understand a real example. Strong judgment can appear in a failed project; a successful result alone does not prove personal causation.
The labels beside capability references describe that capability’s role in this standard; they do not make the whole standard optional. Any role or scenario condition is stated separately.
Evidence to inspect: Reviews adoption, completion, or user feedback after release and adjusts.
Evidence to inspect: Explains the choice, alternatives, and conditions that could change the result.
Applies when: Do not assess by AI usage frequency or speed; consider task suitability and available validation.
Evidence to inspect: Compares workflow time or quality and retains human review points.
Evidence to inspect: Plans include outcome hypotheses, dependencies, and changes.
Evidence to inspect: Design supports requirements and explains capacity, migration, or rollback.
Evidence to inspect: Tests or runtime signals demonstrate that key constraints are protected.
Applies when: Assess only when AI is actually used or the role is expected to evaluate AI output.
Evidence to inspect: Evidence includes tests, comparison cases, reference checks, or human approval.
Evidence to inspect: The model explains exceptions, historical data, and migration checks.
Evidence to inspect: Release monitoring, rollback conditions, and verified repairs form a closed loop.
Evidence to inspect: Decisions, dependencies, and owners are clear; changes are shared promptly.
Evidence to inspect: Others reuse the practice; quality issues or handoff costs decline.
Evidence to inspect: The disagreement ends with a decision, owner, and review point.
Evidence to inspect: A peer completes a change or diagnosis using the handoff material.
Evidence to inspect: Shows how new evidence changed priority or implementation.
Evidence to inspect: Partners can plan around updates and face fewer surprises.
Applies when: Allows a context-based decision not to use AI.
Evidence to inspect: Identifies data, access, accuracy, or impact-scope risks.
Evidence to inspect: Clarifying questions change the problem framing or investigation, with explicit limits to inference.
Evidence to inspect: Explains concrete changes after failure and the next attempt's results.
