Data engineering · Stage expectations

Check grain, joins, and calculations with guidance and retain reproduction steps.

Evidence：Metric definitions, reproducible analyses, lineage, quality checks, or model evaluations Definitions, grain, sample bias, temporal leakage, privacy, and decision risk

Own personal actions, checks, and timely requests for help on defined tasks. Represent data objects, grain, time, lineage, and definitions and check their correspondence with real events.

Evidence：Metric definitions, reproducible analyses, lineage, quality checks, or model evaluations Definitions, grain, sample bias, temporal leakage, privacy, and decision risk

Complete a bounded data task, explain definitions, and check anomalies. Explain the basic approach and recognize a feasible alternative and its conditions.

Evidence：Personal steps, before/after differences, checks, and changes after feedback.

Own personal actions, checks, and timely requests for help on defined tasks. Check task completion, report errors and unknowns, and do not claim the whole team result.

Evidence：Personal steps, before/after differences, checks, and changes after feedback.

Keep collaborators informed about progress and blockers and confirm feedback was applied.

Evidence：Personal steps, before/after differences, checks, and changes after feedback.

People management is not a general requirement at this stage. Recognize responsibilities and ask for help; do not require hiring, performance reviews, or organization design.

Evidence：Personal steps, before/after differences, checks, and changes after feedback.

Explain how new knowledge was applied and how understanding was checked. Separate facts, guesses, and unknowns, restate the problem, and check basic causal claims. Seek necessary information, complete checks, and ask specific questions when blocked. Explain an adjustment after an error or setback; resilience does not mean accepting unreasonable pressure.

Evidence：Personal steps, before/after differences, checks, and changes after feedback.

Understand responsibility, privacy, and escalation boundaries; hiring, performance management, and organization design are not expected. Integrity, safety, and truthful attribution apply at every stage; depth distinguishes the scope of prevention and correction.

Evidence：Personal steps, before/after differences, checks, and changes after feedback.

Select fitting methods, validate quality and bias, and distinguish correlation from causation.

Evidence：Metric definitions, reproducible analyses, lineage, quality checks, or model evaluations Definitions, grain, sample bias, temporal leakage, privacy, and decision risk

Own routine tasks or a component from clarification through validation. Represent data objects, grain, time, lineage, and definitions and check their correspondence with real events.

Evidence：Metric definitions, reproducible analyses, lineage, quality checks, or model evaluations Definitions, grain, sample bias, temporal leakage, privacy, and decision risk

Own an analysis, pipeline, or model task and explain the decision it supports. Compare feasible options under goals and constraints and explain the choice and rejected alternatives.

Evidence：Independent judgment, option records, validation definitions, collaboration interfaces, and closed-loop outcomes.

Own routine tasks or a component from clarification through validation. Explain a comparable baseline, outcome definition, and observation window and separate personal action from shared output.

Evidence：Independent judgment, option records, validation definitions, collaboration interfaces, and closed-loop outcomes.

Coordinate direct dependencies, clarify commitments, and help collaborators deliver shared work.

Evidence：Independent judgment, option records, validation definitions, collaboration interfaces, and closed-loop outcomes.

Support peers, mentoring, and professional ownership within the mandate; broader influence does not imply authority over hiring or performance.

Evidence：Independent judgment, option records, validation definitions, collaboration interfaces, and closed-loop outcomes.

Explain how learning sources were selected, application verified, and unsuitable experience corrected. Break goals into testable hypotheses, compare evidence, and explain the conditions supporting a conclusion. Identify gaps in closure and arrange validation and follow-up beyond assigned steps. Review failed assumptions and recovery and show what changed in the next task.

Evidence：Independent judgment, option records, validation definitions, collaboration interfaces, and closed-loop outcomes.

Decide within authority and escalate cross-boundary or sustained risks; peer support is not full people management. Integrity, safety, and truthful attribution apply at every stage; depth distinguishes the scope of prevention and correction.

Evidence：Independent judgment, option records, validation definitions, collaboration interfaces, and closed-loop outcomes.

Compare analysis or modeling approaches and handle time semantics, data changes, and validation bias.

Evidence：Metric definitions, reproducible analyses, lineage, quality checks, or model evaluations Definitions, grain, sample bias, temporal leakage, privacy, and decision risk

Own complex or ambiguous problems, coordinate key dependencies, and enable others to own work. Represent data objects, grain, time, lineage, and definitions and check their correspondence with real events.

Evidence：Metric definitions, reproducible analyses, lineage, quality checks, or model evaluations Definitions, grain, sample bias, temporal leakage, privacy, and decision risk

Resolve complex data or definition conflicts and verify effects on decisions and production. Compare plausible paths, opportunity costs, and counterexamples and explain fit to current constraints.

Evidence：Complex constraints, decisions, critical personal actions, counterevidence, and changes in later reuse.

Own complex or ambiguous problems, coordinate key dependencies, and enable others to own work. Validate outcomes, adverse effects, and alternative explanations and extract transferable but bounded insights.

Evidence：Complex constraints, decisions, critical personal actions, counterevidence, and changes in later reuse.

Resolve cross-role disagreement, move decisions forward, and expand impact through review, coaching, or shared practice.

Evidence：Complex constraints, decisions, critical personal actions, counterevidence, and changes in later reuse.

Support peers, mentoring, and professional ownership within the mandate; broader influence does not imply authority over hiring or performance.

Evidence：Complex constraints, decisions, critical personal actions, counterevidence, and changes in later reuse.

Transfer methods across contexts and revise judgment using counterevidence. Structure ambiguous problems, test alternative explanations, and identify unknowns that change decisions. Identify consequential problems and establish validation cadence rather than simply increase workload. Adjust goals, methods, or collaboration after setbacks and verify sustained improvement.

Evidence：Complex constraints, decisions, critical personal actions, counterevidence, and changes in later reuse.

Explain coordination authority and escalation paths; mentoring and craft influence do not imply personnel authority. Integrity, safety, and truthful attribution apply at every stage; depth distinguishes the scope of prevention and correction.

Evidence：Complex constraints, decisions, critical personal actions, counterevidence, and changes in later reuse.

Establish reusable data standards, methods, and evolution practices and validate adoption.

Evidence：Metric definitions, reproducible analyses, lineage, quality checks, or model evaluations Definitions, grain, sample bias, temporal leakage, privacy, and decision risk

Own a critical professional domain and scale impact through judgment and shared practice, not direct-report count. Represent data objects, grain, time, lineage, and definitions and check their correspondence with real events.

Evidence：Metric definitions, reproducible analyses, lineage, quality checks, or model evaluations Definitions, grain, sample bias, temporal leakage, privacy, and decision risk

Own methods, trustworthiness, and sustained adoption in a critical data domain. Compare long-term directions, domain alternatives, and migration costs; establish internal baselines when external comparisons are unavailable.

Evidence：Domain baselines, critical judgments, adoption and transfer records, long-term outcomes, and attribution limits.

Own a critical professional domain and scale impact through judgment and shared practice, not direct-report count. Check sustained adoption, cross-context effects, and costs and explain the distinctive personal judgment.

Evidence：Domain baselines, critical judgments, adoption and transfer records, long-term outcomes, and attribution limits.

Build sustained influence through cross-team mechanisms, professional standards, and developing owners.

Evidence：Domain baselines, critical judgments, adoption and transfer records, long-term outcomes, and attribution limits.

Support peers, mentoring, and professional ownership within the mandate; broader influence does not imply authority over hiring or performance.

Evidence：Domain baselines, critical judgments, adoption and transfer records, long-term outcomes, and attribution limits.

Renew domain methods and explain when prior experience fails and how new directions are validated. Develop domain models, recognize local optima and system effects, and explain limits of method transfer. Identify consequential long-term problems, advance necessary unglamorous work, and set stopping or revision criteria. Correct consequential judgment errors openly and help affected teams recover and learn.

Evidence：Domain baselines, critical judgments, adoption and transfer records, long-term outcomes, and attribution limits.

Make authority and organizational dependencies explicit and do not claim environmental advantages as personal achievement. Integrity, safety, and truthful attribution apply at every stage; depth distinguishes the scope of prevention and correction.

Evidence：Domain baselines, critical judgments, adoption and transfer records, long-term outcomes, and attribution limits.

Align data ownership, cross-domain definitions, and governance responsibilities while developing leaders.

Evidence：Metric definitions, reproducible analyses, lineage, quality checks, or model evaluations Definitions, grain, sample bias, temporal leakage, privacy, and decision risk

Own multi-team or organizational capability through managers with explicit portfolio and decision rights. Represent data objects, grain, time, lineage, and definitions and check their correspondence with real events.

Evidence：Metric definitions, reproducible analyses, lineage, quality checks, or model evaluations Definitions, grain, sample bias, temporal leakage, privacy, and decision risk

Allocate data capability through multiple leaders and validate portfolio value. Compare portfolios, organization options, resource flows, and long-term capability and examine stakeholder interests and constraints.

Evidence：Organization/resource alternatives, manager development, succession coverage, multi-cycle outcomes, counterevidence, and correction mechanisms.

Own multi-team or organizational capability through managers with explicit portfolio and decision rights. Inspect organizational outcomes across cycles, management quality, succession, and adverse effects and separate mechanisms from external conditions.

Evidence：Organization/resource alternatives, manager development, succession coverage, multi-cycle outcomes, counterevidence, and correction mechanisms.

Develop, assess, and empower managers and calibrate organizational boundaries, management quality, and succession.

Evidence：Organization/resource alternatives, manager development, succession coverage, multi-cycle outcomes, counterevidence, and correction mechanisms.

Develop, assess, and empower managers and calibrate organizational boundaries, management quality, and succession.

Evidence：Organization/resource alternatives, manager development, succession coverage, multi-cycle outcomes, counterevidence, and correction mechanisms.

Turn cross-team reviews and external changes into organizational practice and check actual adoption. Identify organizational feedback loops, incentives, and delays and use counterevidence to revise strategic and organizational assumptions. Anticipate capability gaps, build leadership pipelines, and resource long-term problems. Own correction during crises and strategic errors and preserve truthful information and organizational recovery.

Evidence：Organization/resource alternatives, manager development, succession coverage, multi-cycle outcomes, counterevidence, and correction mechanisms.

Establish upward information flow, conflict-of-interest controls, and decision review rather than routine bypassing. Integrity, safety, and truthful attribution apply at every stage; depth distinguishes the scope of prevention and correction.

Evidence：Organization/resource alternatives, manager development, succession coverage, multi-cycle outcomes, counterevidence, and correction mechanisms.