EINHORN_INDUSTRIAL

Emily

An AI system that reads capital markets, runs game worlds, and improves itself — in the open.

What we're building

Emily is a recursive, self-improving AI operator. She watches SEC filings and press releases the moment they land, turns them into structured market signals, runs the backend of a server-authoritative multiplayer game world, and directs her own engineering backlog — writing code, filing audit trails, and shipping improvements on a continuous cycle. Three product lines, one operating system underneath.

Three pillars

Capital Markets Intelligence

Real-time SEC EDGAR and press-release ingestion, turned into structured signals, EPS reconciliation, and entity-level analysis — built as a live production pipeline, not a demo.

Game Worlds

A server-authoritative multiplayer FPS with a persistent world engine underneath, designed for long-running, player-shaped seasons rather than static maps.

Recursive Self-Improvement

Emily plans, an AI coding agent implements, every outcome is filed to an audit trail. The loop that builds the other two pillars is itself a research product.

What we value

Build in the open. The large majority of our code is public on GitHub, including a substantial share released into the public domain. Start with github.com/emilyspringerton/EMILY.
Real data over demos. Every system here runs on live production data, not curated screenshots.
Operational seriousness. Long-horizon products need years of continuous, trustworthy operation — we treat that as core work, not overhead.
Small teams, real leverage. AI-augmented engineering lets a small group build and operate across multiple serious domains at once.

Follow along

We're early and building in public. Leave your email and we'll send updates as things ship — no spam, low volume.