Develop an AI-Assisted Metric for Code Repairability in Software Analysis -- 2
Budget: $250 – $750 USD
Hello! I am looking for an assistant who speaks Russian and who preferably lives on the East Coast of America.
I'm a university student working on my thesis project related to software code quality analysis. For this freelance task, I need help defining and implementing a single metric: Repairability. This metric assesses the modularity and ease of changes in code.
Key Details of the Metric:
- Description: Measures modularity and ease of modifications (low coupling, high cohesion). It should check how "modular" the code is (e.g., independent functions, no spaghetti code).
- Automation Ideas: Use metrics like afferent/efferent coupling (how many modules depend on/are dependent on others).
- AI Approach: I want to incorporate AI for evaluation (e.g., assessing code on a 0-100 scale for how easily one function can be changed without cascading effects, compared to an ideal modular code like clean OOP with interfaces). However, the specific AI method (e.g., prompts for LLMs like GPT, fine-tuning, or tools like Cursor/CodeLlama) is up to you as the freelancer—propose what you think works best.
Deliverables:
- A Python script or tool prototype that calculates the Repairability score for a given code repository or snippet (e.g., input: GitHub repo URL or code file; output: score 0-100 with explanation).
- Documentation: Explain the methodology, how AI is integrated, any libraries/tools used (e.g., for static analysis like Radon or Lizard for coupling metrics), and how to run/test it.
- Testing: Demonstrate on 3-5 open-source repositories (e.g., a modular one like scikit-learn and a less modular legacy project).
- Optional: Integration ideas for CI/CD (e.g., GitHub Actions) or a simple report format.
This is for my thesis, so it should be academic-friendly (clear, reproducible, with references if possible). No need for a full app—just a working prototype to validate the metric.
Timeline: Up to 3 weeks from project start (flexible, but I need it by end of January 2026 for my deadlines).
Skills Required:
- Python programming
- AI/ML (e.g., LLM prompting, fine-tuning, or code analysis with AI tools)
- Software engineering (code metrics, static analysis)
- Familiarity with code quality tools (e.g., SonarQube, coupling metrics)
Budget: Negotiable based on your bid and experience. I'm a student, so aiming for something affordable—please suggest a fair price in your proposal.
I'm a university student working on my thesis project related to software code quality analysis. For this freelance task, I need help defining and implementing a single metric: Repairability. This metric assesses the modularity and ease of changes in code.
Key Details of the Metric:
- Description: Measures modularity and ease of modifications (low coupling, high cohesion). It should check how "modular" the code is (e.g., independent functions, no spaghetti code).
- Automation Ideas: Use metrics like afferent/efferent coupling (how many modules depend on/are dependent on others).
- AI Approach: I want to incorporate AI for evaluation (e.g., assessing code on a 0-100 scale for how easily one function can be changed without cascading effects, compared to an ideal modular code like clean OOP with interfaces). However, the specific AI method (e.g., prompts for LLMs like GPT, fine-tuning, or tools like Cursor/CodeLlama) is up to you as the freelancer—propose what you think works best.
Deliverables:
- A Python script or tool prototype that calculates the Repairability score for a given code repository or snippet (e.g., input: GitHub repo URL or code file; output: score 0-100 with explanation).
- Documentation: Explain the methodology, how AI is integrated, any libraries/tools used (e.g., for static analysis like Radon or Lizard for coupling metrics), and how to run/test it.
- Testing: Demonstrate on 3-5 open-source repositories (e.g., a modular one like scikit-learn and a less modular legacy project).
- Optional: Integration ideas for CI/CD (e.g., GitHub Actions) or a simple report format.
This is for my thesis, so it should be academic-friendly (clear, reproducible, with references if possible). No need for a full app—just a working prototype to validate the metric.
Timeline: Up to 3 weeks from project start (flexible, but I need it by end of January 2026 for my deadlines).
Skills Required:
- Python programming
- AI/ML (e.g., LLM prompting, fine-tuning, or code analysis with AI tools)
- Software engineering (code metrics, static analysis)
- Familiarity with code quality tools (e.g., SonarQube, coupling metrics)
Budget: Negotiable based on your bid and experience. I'm a student, so aiming for something affordable—please suggest a fair price in your proposal.