qapton

A closed research team studying how to predict financial markets across world indices, US, European and Asian equities, and crypto.

Research Scope

Markets

  • World indices
  • US equities
  • European equities
  • Asian equities
  • Other major regions

Prediction Targets

  • Equities
  • Futures
  • Options
  • Auctions
  • Crypto tokens

Research Streams

AI tooling for prediction research
Builds and evaluates RAG systems, MCP servers, and small language models (SLMs) to measure their effect on prediction quality.
Prompt-engineering effects on LLM predictions
Investigates how prompt-engineering techniques alter prediction outcomes from LLMs — including adversarial and unconventional prompting patterns.
Multi-agent LLM session effects
Studies how multi-agent orchestration patterns within LLM sessions affect prediction quality.
Quantitative trading systems
Builds quantitative trading systems that execute automated trading decisions on top of the team's prediction work.
Backtesting systems
Builds backtesting infrastructure to evaluate prediction models and trading strategies against historical data.
Non-standard data sources
Collects and processes non-standard data sources for use as prediction inputs — beyond conventional market-data feeds.
LLM self-learning and feedback
Investigates self-learning mechanisms and feedback loops in LLMs, and how they affect prediction quality.

Operating Model

Research findings, methodologies, insights, and accumulated experience are not shared, sold, or published externally. The team is a closed research operation — closure is part of the identity, not an interim posture.

Posture

Closed research team with no external publication or commercialization.

Data Sharing

Research findings and methodologies remain internal to the team.