AI Workbench

From daily agents to AI-native ventures.

A practical stack for research, coding, operations, and product prototypes.

Daily Stack

Tools arranged by how they actually enter the workflow.

CC
Coding Agent

Claude Code

Anthropic's agentic coding environment — my primary dev tool. Full-stack development, debugging complex TypeScript/Python projects, and building this portfolio. Acts as a pair programmer that understands the entire codebase.

OA
Reasoning & Code

GPT o3 / Codex

OpenAI's advanced reasoning models. Used for complex econometric problem-solving, multi-step policy analysis requiring deep logical chains, and structuring long-form research outlines.

MM
日常运营 / Daily Ops

Minimax 2.7

Integrated via OpenClaw for daily workflows — document summarization, Chinese-English translation, meeting notes, and rapid literature synthesis. Extended context window ideal for lengthy policy documents.

OC
AI Deployment

OpenClaw

AI orchestration platform for deploying and managing multiple model APIs. Routes tasks to the optimal model, manages context windows, and builds automation pipelines for research workflows.

CU
Dev Environment

Cursor IDE

AI-native code editor built on VS Code. Inline AI completions and codebase-aware chat dramatically accelerate development of data analysis scripts, web apps, and automation tools.

CZ
Agent Builder

Coze Platform

Low-code platform for building conversational AI agents. Used to build the Dashboard AI Assistant — configuring RAG workflows, intent classification, and DeepSeek-V3.1 model integration for academic advising.

PP
Research

Perplexity Pro

Real-time AI search for academic and policy research. Rapidly surveys literature, tracks breaking policy developments, and cross-references sources — especially useful for fast-changing GBA policy research.

ST
Econometrics

Stata + AI

Traditional econometric workflows in Stata combined with AI-assisted code generation and output interpretation. Claude writes regression code and explains results — dramatically accelerating empirical policy research.

AI Ventures

Product ideas moving from coursework into systems.