AI · Quant · System
AI Quant Research Platform
An agentic research and trading system for market analysis, signal generation, strategy testing, risk management, and automated execution.
- My role
- Builder — research workflow design, agent tooling and risk-control logic.
- Status
- Research and build in progress.
What it does
A research environment where agents gather market and news data, propose signals, test strategies against historical data, and surface results with the risk constraints attached — with execution kept behind explicit limits.
The problem
Quant research produces a lot of promising-looking output and very little of it survives honest testing. The gap is rarely modelling ability; it is disciplined data handling, testing and risk control.
The approach
Separate stages for data collection, signal generation, strategy testing and risk evaluation, with agents assisting each stage rather than producing an end-to-end answer. Risk limits are part of the system, not a setting applied afterwards.
- Strict separation between research and execution paths.
- Testing harness first, so a strategy has to be falsifiable before it is interesting.
- Hard risk limits enforced in code rather than in prompts.
- Agents used for analysis and hypothesis generation, not for unchecked decisions.
Outcome & status
Research and build in progress; no performance claims are made and none should be inferred.