Bhuwan.dev
Back to Projects

Repause

AI-driven repository analysis for extracting technical narratives and interview prep.

🟢 LiveAI

Project Overview

Repause is an AI-powered project defense platform that transforms GitHub repositories into personalized interview preparation material. By analyzing a repository's codebase, project structure, dependencies, and technologies, it generates technical narratives, architectural explanations, and project-specific interview questions. Users can practice defending their engineering decisions, understand architectural tradeoffs, and improve their technical communication using their own projects as the source of truth.

Why I Built This

Many developers can build impressive projects but struggle to explain why they made certain technical decisions during interviews. Recruiters and interviewers often evaluate a candidate's ability to justify architecture, discuss tradeoffs, and communicate engineering decisions—not just write code. Existing AI interview platforms focus primarily on coding questions or behavioral interviews rather than helping candidates articulate their own projects.

Architecture & Core Components

The platform follows a modular AI pipeline designed for repository analysis and narrative generation.

GitHub Integration

Authenticates users and imports repositories through the GitHub API.

Repository Analyzer

Extracts frameworks, dependencies, folder structure, APIs, databases, and architectural patterns.

AI Reasoning Engine

Uses LLMs to infer design decisions, identify tradeoffs, and generate technical explanations.

Interview Simulator

Creates project-specific questions and enables candidates to practice articulating their decisions.

Tradeoffs

GitHub as primary source

Benefit: Uses real code instead of manual descriptions.
Tradeoff: Limited insight into undocumented business decisions.

AI-generated narratives

Benefit: Personalized interview prep with minimal effort.
Tradeoff: May occasionally infer decisions that require user verification.

Modular analysis pipeline

Benefit: Easier to extend with new languages, frameworks, and LLMs.
Tradeoff: Slightly higher architectural complexity.

LLM-based reasoning

Benefit: Produces contextual and natural explanations.
Tradeoff: Higher inference cost and latency compared to rule-based systems.

Engineering Highlights

AI-powered GitHub Analysis

Automated repository parsing and metadata extraction using modular pipelines.

Technical Narrative Generation

Generates readable architecture and engineering decision explanations.

Personalized Interview Prep

Dynamic generation of project-specific technical questions for robust practice.

Tradeoff Articulation

Empowers users by clearly presenting and evaluating complex engineering decisions.