QuantSilico

QuantSilico is organised as connected product surfaces rather than a single sequence. An agent and automation layer handles research tasks, evidence collection, comparison and report generation, and it assists research without bypassing governance. The core research surfaces are Studio (strategy specifications, universes, reusable components, configuration), Discovery (candidate search, comparisons, research prompts, agent-assisted investigation and idea critique) and Validator (leakage checks, cost assumptions, walk-forward tests, sensitivity, stability and promotion evidence). The Validator gates promotion into a governance layer of approval gates, risk limits, audit trail, access boundaries and deployment state. Only after approval do the execution surfaces run: Deploy handles approved configuration bundles, paper and live adapters, promotion controls and execution handoff, and Cockpit shows portfolio state, risk, telemetry, alerts, intervention and post-run review. A Ledger acts as a memory backbone beneath the whole system, recording datasets, versions, experiments, configurations, metrics, results, decisions and approvals; every surface writes evidence to it, and Cockpit telemetry feeds back as evidence. This is the intended product direction, not a claim that each surface is complete.