Skip to content

Repository files navigation

Interview Journey

Interview Journey is an evidence-driven interview preparation system. It prepares candidates for technical hiring processes using every piece of available interview information — the target role, the candidate's real resume evidence, the current interview stage, company and team context, previous recruiter conversations, previous interview questions, previous feedback, technical assessments, interviewer information, candidate observations, and lessons from earlier stages.

Problem

Generic interview prep ("here are 50 common questions") ignores what actually matters for a specific candidate in a specific hiring process at a specific stage. There is no lightweight, repeatable method for turning a target role, a real resume, and scattered interview signals into a focused plan for the very next interview.

Intended users

Individual candidates preparing for technical interviews — engineers at any seniority — who want a structured, on-demand preparation assistant that adapts to their real background and their actual hiring process, rather than a generic study guide.

Planned product surfaces

  • Claude Skill — an installable Skill adapting the shared methodology for execution inside Claude.
  • Claude Project — a guided conversational interface over the Skill and shared Knowledge.
  • ChatGPT Custom GPT — a conversational interface to the same methodology, built for the ChatGPT ecosystem.

All three surfaces are thin, platform-specific wrappers. The methodology itself lives once, in frameworks/ and core/.

Core architectural principle

Shared methodology (frameworks/, core/, schemas/, workflows/, outputs/)
        ↓
Platform-independent core
        ↓
Claude Skill / Claude Project / ChatGPT GPT

core/ and frameworks/ are the canonical source of product behavior. Platform folders (claude/, chatgpt/) only adapt that shared definition to a specific product surface; they must never redefine or duplicate business rules. This repository reuses the architectural pattern of Career Targeting Intelligence — canonical methodology, platform adapters, explicit packaging allowlists, deterministic cross-platform scripts — while its actual content is built around technical interview preparation, not career/company targeting.

Repository structure

interview-journey/
├── core/            platform-independent product definition, policy, and workflow
├── frameworks/       01–15, the canonical methodology (15 is the master orchestrator)
├── schemas/          structured records for every framework's output
├── workflows/        orchestration and task-specific flow adapters
├── outputs/          canonical output contract templates
├── claude/           Claude Skill, Claude Project, and external-install packaging
├── chatgpt/          ChatGPT Custom GPT instructions, Knowledge, and packaging
├── scripts/          deterministic packaging/build scripts (Bash + PowerShell)
├── examples/         a fully synthetic worked example
├── tests/            repository validation
└── dist/             generated packages (git-ignored)

Framework model

Fifteen frameworks define the methodology, each with a single responsibility:

# Framework Responsibility
01 Role Intelligence Structured hiring-intent analysis of a target role
02 Resume Intelligence What the candidate can actually demonstrate
03 Interview Stage Where the candidate is in the hiring process
04 Role Fit & Gap Analysis Role Intelligence vs. Resume Intelligence
05 Interview Intelligence Continuous enrichment from new interview evidence
06 Preparation Strategy The shortest, highest-impact preparation plan
07 Question Prediction Evidence-based likely questions
08 Interview Hypothesis What the interviewer is likely trying to validate
09 Coding Interview Decision Engine Coding preparation, simulation, and evaluation
10 System Design Framework System-design preparation, simulation, and review
11 Behavioral Interview Framework Real-experience-based behavioral preparation
12 Mock Interview Framework Full realistic interview simulation
13 Answer Coaching Framework Structured feedback on a submitted answer
14 Post-Interview Debrief Framework Structured learning from a completed interview
15 Interview Journey Intelligence Framework Master orchestrator — decides which frameworks run, in what order

Document 15 does not replace 01–14; it decides how they work together. See core/orchestration-policy.md.

Workflow routing

State handling

core/state-management.md and schemas/interview-journey-state.schema.md define the Interview Journey State model. State is a logical record, not a storage mechanism: no background automation or hidden persistent data storage is implemented or implied. State is represented explicitly through user-provided files, Project Knowledge, conversation context, or generated artifacts.

Claude Skill

./scripts/package-claude-skill.sh
.\scripts\package-claude-skill.ps1

Both produce dist/interview-journey.skill.zip.

Claude Project experience

For teammates outside the creator's Claude organization, a portable external self-install kit packages the Skill, Instructions, and approved Knowledge into dist/interview-journey-claude-kit.zip, built with ./scripts/package-claude-external-kit.sh or .\scripts\package-claude-external-kit.ps1.

ChatGPT Custom GPT

./scripts/build-chatgpt-knowledge.sh
./scripts/package-chatgpt-gpt.sh
.\scripts\build-chatgpt-knowledge.ps1
.\scripts\package-chatgpt-gpt.ps1

Both produce dist/interview-journey-chatgpt.zip. This package configures no Actions, Apps, or external APIs.

Synthetic example

A complete, fully fictional worked example — candidate, company, recruiter, interviewers, questions, and feedback — demonstrates the frameworks, schemas, and output contracts working together end to end:

Limitations

  • No background monitoring, no scheduled research, no automatic persistence.
  • No ChatGPT Actions, no Claude MCP servers, no external APIs, no databases.
  • The product cannot guarantee cross-conversation memory beyond what the underlying platform itself retains.
  • Missing source frameworks are never fabricated — see core/evidence-policy.md.

Privacy boundaries

The product may only analyze information explicitly provided by the user. It must never invent candidate experience, interview process details, or interviewer information, and it must never copy real personal data into shared repository files, Skill packages, or Knowledge bundles — see core/state-management.md#context-boundary.

Further reading

About

Evidence-driven interview preparation for any role, across ChatGPT and Claude.

Topics

Resources

Stars

0 stars

Watchers

0 watching

Forks

Releases

Packages

Contributors

Languages