A.10 Job Search Preparation Checklist

Job preparation is not “learn everything first.” It is turning learning traces into projects other people can understand.

Pick one target role first

Role	What matters most	Prepare
AI / Algorithm Engineer	Model understanding, training, evaluation	ML/DL projects, metrics, experiments
LLM Application Engineer	RAG, Agent, backend, product loop	Complete app, API design, logs, evaluation
Data Analyst / Data Scientist	SQL, statistics, visualization, modeling	Analysis reports and business explanations
AI Product / Technical Product	Scenario judgment, requirements, evaluation	Product proposal, metrics, trade-offs

Do not prepare for every direction at once.

Project story format

Use this structure in resume, README, and interviews:

Field	What to write
Target role	The role this project is meant to support
User problem	The concrete problem or workflow you improved
Input and output	What goes in, what comes out, and who uses it
Baseline	The simplest comparison point or previous workflow
Technical solution	The main system design, model, data, or product choice
Evaluation result	The metric, test set, user check, or reproducible evidence
Failure case	One thing that did not work or remains risky
What I improved	The specific change you made after seeing evidence

Weak:

Used Python and LangChain to build a knowledge base Q&A system.

Stronger:

Built an enterprise knowledge base Q&A system with document chunking, vector retrieval,
permission filtering, and cited answers; created an evaluation set to compare chunking
strategies and reduce false retrievals.

Repository checklist

• README
• How to run
• Project structure
• Example input and output
• Screenshots or demo images
• Metrics or evaluation method
• Known issues and next steps

Someone opening the repo should understand the project in 3 minutes.

Interview questions to prepare

• Why did you choose this solution?
• What baseline did you compare against?
• What failed?
• How did you evaluate the result?
• If production breaks, what do you check first?

Four-week sprint

Week	Focus
1	Choose target role and select 2-3 projects
2	Improve README, screenshots, run instructions, resume wording
3	Practice project explanation and fundamentals
4	Apply, record questions, improve projects from feedback

Evidence to Keep

Keep this page’s proof of learning as a small evidence card:

Target Role

AI full-stack, LLM app, data/ML, Agent engineer, or multimodal builder

Portfolio Story

problem, system, evidence, failure, improvement, and trade-off

Gap List

missing project, weak explanation, missing deployment, or unclear metrics

Next Action

one resume/project/interview artifact to update this week

Expected Output

a portfolio story card that can be used in an interview