Scientific publications

We are part of the academic world, and we actively publish our research.

The strands cover the two axes of our work: quantitative finance and LLM agent engineering. Every paper is public and verifiable: read, replicate, criticise.

The six papers

  • Streamlined Hierarchical Reinforcement Learning for Algorithmic Trading: Architecture Simplification and Empirical Validation

    Quant / ML · October 2025 · 42 pp.

  • Feature Scope and Cross-Sectional Return Prediction: Evidence from US Large-Cap Equities

    Quant / ML · April 2026 · 98 pp.

  • Conformance, Cost, and Replication in Constrained LLM Coding Agents

    LLM agents · May 2026 · 32 pp.

  • Beyond Retrieval vs Context: A Unified Evaluation Framework for External Information Management in LLM Agents

    LLM agents · June 2026 · 75 pp.

  • Who Captures Atomic MEV after PBS? Receipt-exact Evidence of Builder Centralization and a Vanishing Independent-searcher Edge on BNB Smart Chain

    Blockchain / MEV · July 2026 · 37 pp.

  • FLEX-MoE: Failure-Born Orthogonal Experts for Self-Improving Language Models

    LLM agents · July 2026 · 22 pp.

Research is not a showcase: it is the foundation of the method. The quant papers underpin the platforms; the agent papers underpin agentic-sdlc, the way we deliver software.