23/08/2026
Big milestone: Submitting to top Springer-Nature journal in mathematical finance. Built upon decades of work with three GOATs:
• Sandy Grossman — Clark Medalist, co-wrote the information-efficient markets bible with Stiglitz, founded QFS (RenTec/DE Shaw tier). I coded up his 1990 Econometrica simulations.
• Nicos Christofides — my PhD advisor and literal father of graph theory. His TSP algorithm has remained undefeated for over 40 years.
• Tony Constantinides — company science advisor, invented LPNN, co-created JPEG, and was the most-cited IEEE author back-to-back.
The outcome? A unified pipeline that does it all: 1. Info-theoretic profitability bounds 2. TSP combinatorial optimization for bond selection 3. Sub-second LPNN inference 4. Real-time regime detection. All wrapped in 40 pages of bound proofs and empirical validation. Alpha signals? IP-protected. Preprints available to institutional fixed-income investors.
Why this won't become LTCM 2.0:
1. We partner with investors holding existing bond inventories — no liquidity squeezes.
2. Information theory tells you exactly what's needed to hit performance targets, not a game of pure leverage.
3. It runs on fully automated AI infrastructure that does not hallucinate; far more likely to pass compliance.
The technique will also be extended to other FICC products. Full suite live demo at FTAHK, September. DMs open for preprints.
