Quant Dashboard

A small local dashboard for tracking quant interview prep:

• Floating agent for voice-first dashboard updates
• LeetCode completions and review queue
• Green Book chapter progress
• Green Book natural-language study logging
• Quant firm applications
• Live 2027 quant internship sync from northwesternfintech/2027QuantInternships
• Resume upload and open-role fit ranking
• A Codex CLI-backed coach chat with no app-level API key

Run

python3 server.py

Then open:

http://127.0.0.1:8787

Dashboard data is saved in data/dashboard.json.

Agent-First Workflow

Use the floating agent in the bottom-right for normal updates. Tap it to speak. Double tap it to open a small text box. It can route one sentence into multiple actions, for example:

Completed 3Sum, did Green Book Chapter 2 problems 1-12, and submitted SIG QT.

The agent updates the relevant dashboard sections and logs what it changed. The prep and application sections are read-only checklists; the agent is the normal way to update them.

Quant Application Sync

The dashboard auto-syncs live application data from:

https://github.com/northwesternfintech/2027QuantInternships

Imported application rows are shown as a checklist: Done, Not done, or Closed. Any role you mark as Submitted, Applied, OA, Interview, or Offer through the agent stays done across future syncs.

Resume Analysis

Use the Resume Analysis tab to upload a resume. The dashboard extracts the text locally, ranks currently open application rows, and saves a candidate profile with strengths, gaps, target roles, and next actions. The Quant Coach receives that same resume/profile context in every chat.

LeetCode Logging

Tell the floating agent the problem name. The dashboard resolves the title, difficulty, topic, and link, then saves it as completed.

Green Book Logging

Tell the floating agent what you finished. The local Codex-backed parser updates the matching row or creates a new one with chapter, topic, status, completed count, total count, date, and notes.

Codex Chat

The coach panel calls the local Codex executable with read-only project access. The server checks QUANT_CODEX_BIN, codex on PATH, and the macOS app bundle at /Applications/Codex.app/Contents/Resources/codex.

If Codex is somewhere else, set:

QUANT_CODEX_BIN=/path/to/codex python3 server.py

You can also tune the chat timeout:

QUANT_CHAT_TIMEOUT=180 python3 server.py