Research

What we're building

Systems we've built while exploring long-horizon agents, memory, and autonomous research — each one a working loop, not a demo.

Autonomous research

July 2026

ScholarLoop

An autonomous research loop that can't grade its own homework — literature-grounded hypotheses, real experiments, frozen scoring, and papers that cite only verified numbers.

Agents in markets

July 2026

AlphaAgent

A multi-agent platform for markets — ML price models, academic alpha factors, options strategies, and guarded live execution, fused into one auditable decision loop.

Signal & memory

July 2026

LatentFeed

A daily research briefing distilled from the AI firehose — six sources, deduplicated, summarized, and delivered to your notes before you wake up.

More to come — publications and open releases are on the way.