Master student focusing on AI agent engineering, backend systems, and repository-level code automation.

Currently building a coding-agent project inspired by Codex / Claude Code. The project focuses on repository understanding, tool calling, patch generation, pytest-based validation, and failure recovery in real engineering workflows.

• Coding agent workflow for real code repositories
• Tool calling and local command execution
• Repository search, file editing, and patch generation
• Test-driven feedback loop with pytest
• Python backend engineering and LLM application development

KamaClaude Extension A learning-oriented extension based on KamaClaude, focusing on coding-agent workflows for repository-level tasks.

Main components under development:

• Repository file reading and search
• Tool execution interface
• Code patch generation
• pytest execution and failure feedback
• Trace logging for agent actions
• Basic engineering evaluation tasks

Python · Git · Linux · pytest · FastAPI · Docker · LLM API · Tool Calling

I am preparing for AI application engineering / Agent engineering / backend development internship roles. My current priority is to build a small but complete engineering project that can be explained, tested, and improved continuously.