← The team

Principal profile

Shiloh Trowbridge

Founder of Axiom Point Research. Quant trading infrastructure, market microstructure × machine learning, GPU-accelerated computing, real-time systems, derivatives, execution — with fascinations running from the mathematics of compression to formal proof.

STFounder / 01
Role
Founder — Research & Infrastructure
Focus
Systematic derivatives, market data, GPU research systems
Location
United States (remote)

Shiloh Trowbridge is a quantitative engineer with an unusual amount of range. The through-line is systematic markets: he builds trading systems end to end — from raw exchange feeds to signal research, portfolio construction, risk control, and execution audit — where market microstructure meets machine learning, and where research loops have to be fast enough to interrogate every print the market produces. Seven-plus years of that work have left him with a fixed habit of mind: an edge only counts if it survives friction, adverse paths, and cross-examination.

He founded the practice in 2019 as Alpha Bluewater Associates — systematic trading infrastructure operated with direct P&L accountability — and evolved it into Axiom Point Research: a research operation built around the standard that a track record is a claim, and evidence is what survives cross-examination.

His engineering background runs to enterprise scale: production data infrastructure for financial services built on multi-node, GPU-dense systems — the kind of environment where terabytes are table stakes and latency is a budget, not an aspiration — with the same data discipline carried into life-sciences and medical data problems, where the demands on trust and timeliness rhyme with trading’s.

Underneath the systems work is a mathematical streak: a standing fascination with information theory and the mathematics of compression — how much structure a data stream really contains, and what it costs to keep all of it — alongside long-horizon problems in computational complexity and machine-checked proof, worked less for the destination than for the discipline the terrain imposes.

He is the author of Shannon-Cortex, an open-source long-term memory engine for AI coding agents — episodic, semantic, and procedural stores with a nightly consolidation pass modeled on how the hippocampus consolidates experience during sleep — and writes Macro Musings, an occasional long-form series on markets, policy, and macro structure.

Selected systems

Built, operated, and accountable for.

Described at the level of discipline, not blueprint. Methods, signal logic, sizing, and execution specifics stay inside the shop — that line is deliberate, and it is part of the point.

Market structure

Live market-structure analytics

Real-time analytics that watch how the whole market moves together — and how that structure shifts when regimes change. Everything is validated against realized outcomes, never against how convincing a chart looks.

Risk simulation

Simulation at interrogation scale

Risk simulation made interactive enough to argue with: vast scenario ensembles between question and answer, organized around drawdown, tail behavior, and path dependence rather than averages.

Derivatives

Options systems that respect reality

Derivatives engines that connect ideas to executable reality — liquidity, spread cost, capital lockup, and settlement truth — with diagnostics built to separate profitable hindsight from tradable edge.

Data infrastructure

Replay-grade market history

Historical replay engineered for exactness: byte-faithful, reproducible, and fast enough that testing a hypothesis against all of market history is routine rather than a project.

Current fascinations

The threads he keeps pulling.

01

Information theory & compression

How much structure a market really transmits, and what it costs to keep all of it. The mathematics of compression is where storage engineering quietly becomes research advantage.

02

Complexity & formal proof

Machine-checked mathematics and the hard edges of computational complexity — long-horizon problems, worked for the rigor they demand rather than the odds they offer.

03

Memory architectures for AI

Episodic, semantic, and procedural knowledge stores with sleep-cycle-inspired consolidation — the research program behind Shannon-Cortex.

04

Medicine & the life sciences

The latency-and-trust problems markets pose show up in medicine too. He has carried the same real-time data discipline into life-sciences problems, and keeps a standing interest there.

05

Macro structure & writing

Macro Musings — occasional long-form on how markets, policy, and narrative interact.

06

Creative computation

Generative audio-visual systems and other experiments in engineering as a medium, not just a means.

For the right audience

The best read on how he thinks is how he writes.

Rather than a wall of jargon, the working standard is laid out in the research notes — how performance should be evaluated, why execution gets audited, and where robustness claims have to survive inspection. Allocators, researchers, and engineers working adjacent problems are equally welcome to start a conversation.

Working on something adjacent?

Diligence inquiries, research collaborations, and serious infrastructure conversations are all welcome.