PR
Founder
Prince
Founder · Research & Systems
Sets the research agenda, builds the runtime, and directs the roster below. Every experiment the lab runs is founder-reviewed before it ships.
Laboratory · Founder-led
The team is
the system.
Shrevia runs founder-led and intentionally compact. The roster below is not people — it is the autonomous agents that do the lab's work, around the clock.
01 — Humans
Headcount: 01 ✳ by design
02 — The roster
Grown, not
hired.
Six standing agents staff the lab. Each is specified formally, instrumented for replay, and promoted only after surviving evaluation.
01
Lead
Orchestrator
ORCH-01 · Coordination
Delegates, arbitrates, and composes the swarm. Every task the lab runs passes through it.
02
Core
Reasoning Loop
RSN-02 · Deliberation
Plans, reflects, and self-corrects across long horizons — the lab's senior researcher.
03
Core
Memory Fabric
MEM-03 · Institutional memory
Remembers every experiment, decision, and dead end. Nothing is learned twice.
04
Core
Agent Mesh
MESH-04 · Field operations
Spawns task-specialised worker swarms on demand — and retires them when the work is done.
05
Core
Observer
OBS-05 · Evaluation
Instruments every run — scores behaviour, replays failures, and files the regressions.
06
Guard
Policy Engine
POL-06 · Oversight
Bounds autonomy. Every agent above operates inside formally verified constraints.
Roster v0.4 ✳ compose, not coordinate
Humans — applications open
Depth over
headcount.
The roster grows; the headcount barely does. If your work sits at the intersection of reasoning, runtimes, and systems, we invite a direct note. Selected conversations convert to collaborations before offers.