Live · onboarding in waves

About Promethean

We build the platform so the alpha can ship.

Promethean is a two-sided ML trading platform — institutional-grade infrastructure on one side, a curated marketplace for retail investors on the other. We built it infrastructure-first, because quality compounds and garbage strategies ruin reputations.


The thesis

The institutional stack, opened to everyone.

For two decades, the tools that produce real trading edge — feature stores, reproducible pipelines, model registries, point-in-time-correct backtests — have lived inside a handful of quant funds. Retail got the opposite: slot-machine apps optimized for engagement, and robo-advisors that sell you the market average with a fee on top.

Two-sided marketplaces live or die on the quality of what's sold. For trading strategies, quality means reproducibility, risk transparency, and honest track records — none of which exist in today's retail landscape. So we built the institutional stack first, then put a curated marketplace on top of it. Developers get paid for what they ship. Investors get transparent, audited access to real algorithms.

That ordering is the strategy. Infrastructure quality compounds into a moat that a flashier front-end never can.

Founder

Built by people who've shipped this before.

TSTim Shoaf · Founder & CEO
Tim Shoaf
Founder & CEO

Before Promethean, Tim led the ML Platform at Roblox — building feature stores, training infrastructure, and inference systems at scale for one of the largest real-time platforms in the world.

His career sits at the intersection of HPC, distributed systems, and machine learning. Promethean is the tool he wished existed: one declarative SDK, written in Rust, that owns the entire path from raw data to a deployed, reproducible trading strategy.

He built it in the open — shipping the infrastructure first, talking to developers and investors directly, and onboarding in waves.

Tim isn't alone — Promethean was shaped by a founding crew of engineers and operators. Meet the team →

How we build

Principles, not slogans.

A small set of commitments that shape every decision — from the database we picked to the fee split we publish.

01

Reproducible or it didn't happen

Every run pins code, data, and environment. A track record you can't reproduce is marketing, not evidence.

02

Rust-first, all the way down

Business logic lives in Rust; thin bindings expose it. Correctness and performance are not a later optimization.

03

Aligned incentives

Developers keep ~70% of performance fees. We win when their strategies actually perform — not when users churn.

04

Transparency by default

Honest, point-in-time track records and clear risk disclosure. No survivorship bias, no curve-fit screenshots.

05

Infrastructure before hype

We shipped the feature store, pipelines, and registry before announcing a marketplace. Substance compounds.

06

Build in the open

We talk to developers and investors directly, ship in public, and onboard in waves rather than overpromising a launch.

Where we are

Everything, running.

The full stack is live — the infrastructure is real and running, and the marketplace is open.

Shipped

Ember feature store, server-side

Declarative datasets, reactive pipelines, hybrid BM25 + vector search, tensor data over Arrow Flight, git-style branching. Materializes server-side — not a library.

Shipped

Forge pipelines + Model Registry

@pipeline DAGs with a full reproducibility contract, ClickHouse-backed metrics, hyperparameter sweeps, and immutable versioned models with complete lineage.

Shipped

Remote execution on Ray

Submit a pipeline, the platform packages it, spins up a RayJob, streams logs and events back over WebSocket, and persists run status. End-to-end working.

Shipped

Web app + in-browser dev environment

Dioxus/Rust/WASM app with collaborative in-browser workspaces on managed Kubernetes, live DAG visualization, and IdP-agnostic multi-tenant auth.

Live

Backtesting + paper trading

Point-in-time-correct slippage, realistic fills, and competitive leaderboards. Running today.

Live

Marketplace + execution layer

Algorithm storefront with vetting, the investor app, exchange adapters, risk controls, and profit share — all live.

Founding team

Small team, big surface area.

Promethean is early. We're looking for a few people who want to build institutional-grade systems from the ground up — Rust-first, server-side, no legacy. These are the shapes of roles we're growing into, not a fixed org chart.

Engineering

Distributed Systems / Rust

Execution engine, scheduling, and the runtime that makes "commit" turn into trained, deployed models.

Quant

Quant Research / Strategy

Backtesting fidelity, risk models, and the vetting bar that protects the marketplace's reputation.

Product

Full-Stack / Dioxus

The developer experience and the investor app — the surfaces where the platform meets people.

Interested in building this?

We're not posting a careers page yet — we're talking to people directly. If this is the kind of thing you'd want to help build, email tim@cryptex.ai and tell us what you'd want to own.


Interested?

We're onboarding developers and early investors in waves.

Join the waitlist