Skip to content
The Tribunal

The Tribunal

Autonomous consensus for high-stakes decision environments. Specialized agents, verifiable loops, hardware-aware inference.
Technical Overview System Architecture
Current phasePhase 2
Decision modelKnights -> King
Training loopGGUF + LoRA
Inference targetsCUDA / ROCm

Agentic consensus infrastructure

Not a single model making a guess. A chamber of specialized agents leaving an audit trail.

The Tribunal is a decision-making framework that separates signal generation, observation, consensus, and verification. Financial market data is the proving ground; the product is the orchestration layer for autonomous decisions under uncertainty.

01

Specialized inference

The Knights

Independent agents evaluate disjoint views of the state and commit votes without coordinating with each other.

02

Capital and risk gate

The King

Aggregates votes, portfolio context, and guardrails. The King can choose action, scale down, or withhold execution.

03

Silent observer

The Prophet

In Phase 2, Prophet watches and records observations only. Its signal correlation becomes training context later.

Moat-safe public architecture

Show the system. Protect the strategy.

The public surface demonstrates orchestration, inference readiness, and operational maturity without exposing prompts, thresholds, model weights, or proprietary labeling logic.

01State Packetsanitized features
02Knight Votesindependent passes
03King Gateconsensus + risk
04Scribe Ledgerverified feedback

read-only inference stream

Sanitized Consensus Trace
Phase 2
t+00s[INGEST]market-state packet sealed; feature names redacted
t+07s[KNIGHT-IMB]independent vote recorded; confidence band attached
t+14s[KNIGHT-MOM]dissent preserved; no cross-agent coordination
t+21s[PROPHET]observer note stored for Phase 3 training corpus; no vote cast
t+28s[KING]quorum check complete; capital gate remains closed
no prompts exposedno weights exposedaudit trail retained

Infrastructure-first positioning

Built to read as AI systems engineering, not a trading black box.

  • Multi-agent orchestration with role separation and replayable decision logs.
  • Hardware-aware inference across local GPUs and cloud accelerators.
  • Continuous fine-tuning loop where outcomes become supervised training data.
  • Operator-grade watchdogs for pods, artifacts, logs, and cost drift.

Built by Cedrick Baker.

The Tribunal is an agentic orchestration framework for autonomous consensus, where specialized agents deliberate, verify, and leave a record.