Projects

An AI-native paper reading and research platform I designed, built, and operate independently. It has two parts.

The web reader supports paper reading and library management. Agent-generated deep-reading reports help researchers filter papers that do not need a careful read, and understand faster the ones that do.

Paper APIs let local agents (Codex, Claude Code) plug into a user's personal paper library — agents can upload, search, organize, and retrieve papers and reports, and sync selected papers into project workspaces for cross-paper synthesis and literature-review workflows.

Users are primarily PhD students and early-career faculty.

A "cyber mentor" for junior researchers — undergraduates, graduate students, and early-career faculty — built on a structured knowledge base of ~40,000 papers from 30+ top economics and business journals (2000–2026).

Using agentic search — where the agent decides what to search, how many times, and which papers to read in depth across multiple retrieval rounds — it supports:

· Academic writing guidance — distilling writing conventions from top journals
· Econometric method learning — e.g., ask to learn DID and get its development trajectory and representative papers
· Research topic & proposal guidance — finding gaps and sharpening contributions
· Ongoing research mentoring — diagnosing weaknesses in a design and suggesting next steps

The goal is augmentation, not automation — giving every junior researcher access to the kind of guidance that is normally scarce.