salluru.dev
operational

All systems active

SYSTEM ARCHITECTURE
THE REFERENCE STACK
DWG AGT-STACK-01
REV C.3
PROJ SALLURU.DEV
SCALE 1:1 ISO
LAYERS 05
PARTS 15
STATUS ● IN PRODUCTION
FLOW DATA ↑ DECISIONS
OUT-ΣOutput

Autonomous decisions, surfaced and accountable.

  • DecisionsDeterministic decision layer
  • DashboardsOperational telemetry surface
  • HumansHuman oversight gate
◇ select a layer
OUTPUTINTELLIGENCEAGENTSRUNTIMEDATA5-LAYER STACK · 508.0FOOTPRINT 376.0 × 376.0
The anatomy of how I build autonomous systems that survive production.
SCROLL
Reference Architecture

One stack, five layers, end‑to‑end.

This isn't documentation for a single project — it's the operational model I return to whenever I build agentic systems. Raw signal enters at the base; it's processed, executed, orchestrated, reasoned over, and converges to an autonomous decision at the apex — with human oversight wired into the loop. Every layer is designed for the part nobody demos: the seam, the backtest‑to‑live gap, and what happens when something goes wrong at 3am.

5 layers

From raw telemetry at the base to autonomous output at the apex.

15 components

Each a distinct operational part of the running system.

In production

Designed to run continuously, be observed, and make real decisions.

01 / OUT

Output

Autonomous decisions, surfaced and accountable.

DecisionsDeterministic decision layer

The apex — autonomous decisions surfaced to operators, gated by explicit policy.

DashboardsOperational telemetry surface

Real-time views of system state, live signals, and model confidence.

HumansHuman oversight gate

The boundary where a person confirms or overrides before consequential action.

02 / INT

Intelligence

Reasoning, signals, and synthesis.

ModelsFrontier reasoning core

Language-model inference for reasoning, synthesis, and structured, calibrated output.

SignalsRealtime signal engine

Real-time signal processing and multi-source intelligence fusion — separating what matters from the noise.

AnalysisRetrieval + synthesis

Multi-source retrieval and pattern recognition that turns raw data into grounded decisions.

03 / AGT

Agents

Orchestration, workflows, and execution.

OrchestrationMulti-agent coordinator

Routes work across specialized agents with state-machine coordination and clean isolation.

WorkflowsAutomation pipelines

Repeatable automated pipelines for routine, high-volume operations.

AutomationAutonomous execution

Closes the last mile — turning a decision into an action with oversight in the loop.

04 / RT

Runtime

Execution, resilience, and infrastructure.

ExecutionPolyglot runtime

The environment where agent tasks actually run, continuously and observably.

ReliabilityFault-tolerant control

Retries, isolation, circuit-breaking, and graceful degradation at the seams.

InfraEdge / cloud delivery

Deployment and hosting built for systems that run without a babysitter.

05 / DAT

Data

Ingestion, transform, and telemetry.

IngestionRealtime collection

Collecting signals and events from external sources as they happen.

ProcessingStream transform

Real-time normalization and transformation of raw, noisy input.

TelemetryFull-trace observability

Traces, metrics, and logs across every component — visibility before failure cascades.