English · 简体中文
AI Agent Engineering / Java & Python Backend
Computer science undergraduate seeking internships in AI agent and application development.
I build AI applications and developer tools, with a focus on tool calling, context management, and validating generated output. I also explore AI-assisted financial research.
An agent-driven application development workbench. Java manages authentication, per-app tasks, and runtime state; Python DeepAgents / LangGraph handles code generation, incremental edits, and context management. WebSocket delivers live interactions, with Docker-backed execution environments. Existing-project import and verification-driven repair workflows remain under active development.
Java · Spring Boot · Python · DeepAgents · LangGraph · PostgreSQL · Docker
A console built around CLIProxyAPI for gateway lifecycle, model routing, configuration, and usage visibility. My work focuses on the console and client integration; the upstream gateway handles inference forwarding.
A financial research tool exposed through MCP / REST, organizing traceable evidence and calculating returns, volatility, and drawdowns. Currently read-only; it does not execute trades.
Python · MCP · Data Validation · Risk Metrics
Addresses Codex fork-session usage accounting by excluding inherited history from new billable usage, with 11 additional regression tests.
The PR documents before-and-after verification, repeat-scan idempotency, and cross-platform preflight results.
Submitted; not yet merged · Status checked: October 8, 2026
- Execution constraints, context management, and validation for coding agents
- Multi-model client integration and local developer tools
- AI-assisted financial research and evidence validation