AI & Agents in Quant Finance

Machine learning, LLMs, and agent systems on a quant desk.

Machine learning is no longer optional on a quant desk, and language models have turned every analyst into an agent builder. This track moves from learning foundations through LLM mechanics and reasoning patterns, then into the frameworks, retrieval systems, and multi-agent architectures used in practice, closing with honest valuation agents, evaluation and guardrails, and applied AI beyond the trading desk.

Machine Learning Core

ML Foundations for Quants

The statistical-learning toolkit every quant agent is built on, with time-series validation as the discipline that keeps it honest.

Language Models & Agents

LLMs & Agent Basics

An agent is a model in a loop with tools and memory, and its prompt context is the new codebase.

Reasoning Patterns, Critically

Chain-of-thought, self-consistency, ReAct, and tree search buy real reasoning gains, and none of them is verification.

Building Agent Systems

The Agent Framework Landscape

A crowded framework market with one durable core — choose for typed contracts and escape hatches, not loyalty.

Agentic RAG over Filings & News

Grounding answers in EDGAR filings and news, with chunking and citation discipline doing the heavy lifting.

Multi-Agent Architectures

Specialized roles under an orchestrator mirror an investment committee, at a real cost in tokens, latency, and failure modes.

Applied & Production Agents

Fundamental-Analysis Agents

Agents that screen filings and build DCFs must never invent cash flows and must match the cash-flow definition to the discount rate.

Evaluation, Guardrails & Observability

Offline fixtures, calibrated judges, red-teaming, and tracing — evaluation is where model-risk governance meets AI engineering.

AI in Finance Verticals

Claims, fraud, underwriting, and KYC run on the same control stack: auditability, bias testing, drift monitoring, and human override.