Purpose
Most AI-trading interfaces jump straight to execution fantasy instead of building the operator controls, safeguards, and review surfaces that serious use actually requires.
Project · in-progress
TradePilot
A supervised agentic trading assistant with a dashboard-first cockpit, paper-session controls, strategy compilation, and a tamper-evident audit layer.
in-progressFastAPI + DuckDB + Scheduler healthPaper-only trading cockpit built around audit trails, risk checks, and supervised planning.
My role
Designer and engineer
Status
in progress
Methods
- FastAPI
- DuckDB
- Scheduler health
- Paper trading
- Audit ledger
- Hybrid planning
- Operator dashboard
Outputs
- TradePilot combines FastAPI, DuckDB analytics, schedulers, paper-session controls, browser planner fallback, and deterministic risk checks into a GUI-first supervised trading environment.
Result
TradePilot is a systems project about supervision, auditability, and controlled planning under risk, not just market prompts or autonomous hype.
Implementation
How it works
TradePilot combines FastAPI, DuckDB analytics, schedulers, paper-session controls, browser planner fallback, and deterministic risk checks into a GUI-first supervised trading environment.
TradePilot comes out of the tradepilot-prototype workspace, where dashboards, schedulers, paper-session controls, and audit trails matter more than bravado. The dossier keeps that bias intact: this is a review-first trading system built to slow decisions down, not inflate them.