Portfolio Evidence Map

Reusable template for turning learning assets into portfolio and interview evidence. 复用模板：把学习笔记、代码、Dashboard、PRD、评估、架构图和 Capstone 产出转化为求职证据。

Purpose: evidence > claims

The goal is not to say "I understand AI BA / AI PM / AI Architecture". The goal is to show concrete evidence that proves the claim.

目标不是写「我懂 AI BA / AI PM / AI 架构」。 目标是把每个能力主张绑定到可检查、可复述、可面试展开的证据。

Use this map to:

• Select the strongest assets from historical learning work.
• Connect each asset to a role-specific capability.
• Convert raw notes into interview stories and portfolio case studies.
• Identify gaps before applying for AI BA, AI PM, or AI Solutions Architect roles.
• Preserve old learning content while adding clearer evidence layers.

Core principle:

Claim -> Evidence -> Metric / Eval / Control -> Story -> Interview proof

Role positioning matrix

Evidence inventory table

Use one row per asset. Keep raw historical content in place and link to it.

Asset types to include: learning notes, code, dashboards, PRDs, evals, architecture diagrams, controls, and capstone work.

Claim-to-evidence matrix

Write only claims you can prove. Each claim should have at least two evidence links.

Claim quality check: every claim should show your work, include a metric/eval/control/decision, and be explainable in 60 seconds.

Case study narrative builder

Use this when turning one project or learning cluster into a portfolio page.

Suggested title pattern:

[Business problem] + [AI approach] + [measurable outcome]

Example:

Reducing loan service escalation with AI-assisted case triage and controlled human review

Interview proof pack

Prepare this before interviews. The goal is fast recall with evidence.

For each proof pack item, keep:

• One artifact link.
• One metric or eval.
• One tradeoff.
• One failure mode.
• One "what I learned" sentence.

Gap analysis: missing evidence by role

Gap scoring:

• High: role-critical evidence is missing or only theoretical.
• Medium: evidence exists but needs metrics, screenshots, or clearer narrative.
• Low: evidence is usable but can be refreshed or packaged better.

Freshness and versioning rules

Use freshness labels consistently:

Versioning rules:

• Add a "Last reviewed" date when an asset becomes portfolio-ready.
• Add an "Interview angle" note when an asset is selected for applications.
• Add a "2026 relevance" note for older Web3, DeFi, AI, or architecture content.
• Keep original historical notes intact; add updates around them instead of rewriting history.
• Prefer small append-only improvements over large rewrites.

Do-not-delete preservation rule

Do not delete old plans, notes, portfolio assets, or learning artifacts just because they are incomplete, outdated, or not yet polished.

不要删除旧计划、旧笔记、旧作品集、旧学习资产。很多内容尚未学习，后续仍然是长期知识库和证据来源。

When content is outdated:

• Add an update note.
• Add a replacement resource.
• Add a portfolio-readiness status.
• Add a "what changed" section.
• Link it to a newer asset if needed.

Avoid:

• Removing historical learning content.
• Replacing a full old plan with a new plan.
• Rewriting old notes so aggressively that the learning trail disappears.
• Treating old Web3 / DeFi / Solidity / RWA work as irrelevant to AI BA / AI PM / Architecture positioning.

Preservation rule:

Preserve original learning asset -> Add evidence layer -> Add freshness label -> Link to portfolio story

Minimum portfolio release checklist

Before publishing or sharing a portfolio package, confirm:

• At least 1 case study for AI BA positioning.
• At least 1 case study for AI PM positioning.
• At least 1 architecture or ADR-based artifact for AI Solutions Architect positioning.
• At least 1 domain-specific case from financial retail, Web3, DeFi, RWA, risk, compliance, or operations.
• Each case has a clear business problem, stakeholder, workflow, AI capability, metric, and control.
• Each major claim links to at least 2 evidence assets.
• Every portfolio-ready asset has a freshness label.
• Older historical content is preserved and linked rather than deleted.
• Interview proof pack has 5 to 7 themes with 30-second answers.
• At least one artifact shows eval thinking, not only product thinking.
• At least one artifact shows operating model or adoption thinking, not only launch thinking.
• At least one artifact shows architecture tradeoffs, not only a final diagram.
• Sensitive or private data is removed, anonymized, or replaced with synthetic examples.
• Screenshots, dashboards, diagrams, and links are checked before sharing.
• Next improvement is listed for every important gap.

Release note:

Portfolio version:
Target role:
Top 3 proof assets:
Known gaps:
Last reviewed: