Skip to content
Technical Insights

Insights on Financial AI

Technical deep-dives on training data, RL environments, model training, and building AI systems that can operate in financial markets.

01
The Data Problem

The Financial AI Data Problem: Why Market Data Isn't Enough

Financial AI is bottlenecked by a shortage of quality training data. Market data alone doesn't teach models to make decisions.

12 min read
02
Technical Deep-Dive

What is Post-Training for LLMs? A Practical Guide

Post-training transforms base LLMs into useful assistants. Learn how RLHF, DPO, and domain-specific training work.

15 min read
03
Technical Deep-Dive

Anatomy of a Financial Decision Episode

What goes into a complete training episode for financial AI? A technical deep-dive into the data structure.

18 min read
04
The Data Problem

Why Financial AI Needs Reasoning Traces, Not Just Outcomes

Outcomes teach models what happened. Reasoning traces teach models how to think.

10 min read
05
Technical Deep-Dive

Counterfactual Learning: Teaching AI What Could Have Been

Every trade yields one outcome, but multiple learning opportunities. Counterfactual analysis extracts maximum signal.

12 min read
06
Technical Deep-Dive

The Case for Replayable Financial Environments

Static datasets can't train decision-making. Replayable environments let AI explore alternatives.

11 min read
07
Technical Deep-Dive

Human Feedback in Financial AI: Beyond Standard RLHF

OpenAI pays traders $150/hr for feedback. That doesn't scale. Here's what financial AI actually needs.

10 min read
08
Industry Analysis

From Tool Use to Alpha: The Four Stages of Financial AI

Financial AI capabilities evolve in stages. Each stage requires different training data.

14 min read
09
Industry Analysis

Why Traditional Quant ML Isn't the Same as LLM Training

Financial firms have used ML for decades. LLM training is fundamentally different.

11 min read
10
Technical Deep-Dive

Building AI That Can Transact: The Infrastructure Challenge

Getting AI to reliably execute financial transactions requires more than good models.

13 min read
11
Technical Deep-Dive

Social Sentiment for Trading AI: Moving Beyond Headlines

Raw sentiment scores fail. Effective social data requires engagement weighting and noise filtering.

10 min read
12
Industry Analysis

The Continuous Data Problem: Why Financial AI Needs Fresh Training

A financial model trained once becomes stale. Markets evolve and static datasets create brittle systems.

10 min read
13
Industry Analysis

UV Labs vs BloombergGPT vs FinGPT vs Scale AI

Comprehensive comparison of financial AI training data sources for researchers and practitioners.

10 min read