Technical Architecture Visualization
FIG.01 — OVERVIEW
Available for projects

ARCHITECT.BUILD.REVIEW.COMPOUND.

Software production as a closed-loop manufacturing pipeline. Not AI coding - systematic engineering.

47.4724° N, 8.3072° E
V1.0
SPEC-DRIVEN · AGENTIC · EVAL-OPS · GROUNDING · CONTEXT · AUTOMATION · COMPOUND · INFRASTRUCTURE · SPEC-DRIVEN · AGENTIC · EVAL-OPS · GROUNDING · CONTEXT · AUTOMATION · COMPOUND · INFRASTRUCTURE ·
01 THE JOURNEY

Every engineering team goes through the same phases. The question isn't if - it's how fast, and with how much friction.

01
PHASE 1

THE INDIVIDUAL EXPERIMENTS

A developer discovers Copilot. Then Cursor. Maybe Claude. The early wins are impressive - entire functions in minutes. The tool feels like a superpower.

Velocity spike No team standards Everyone uses different tools
02
PHASE 2

THE TEAM ADOPTS

MOST ARE HERE

Management hears about the wins. "Everyone should use this." But what works for individuals doesn't scale. Some devs swear by it, others are skeptical. Code reviews get weird.

Inconsistent adoption Quality debates Velocity plateau
03
PHASE 3

THE PROCESS BREAKS

Fast to write, slow to review. Bugs nobody understands. "It worked when I prompted it." The tool has no context - every prompt starts from zero.

Reviews take longer than writing Same bugs twice Context amnesia
04
PHASE 4

THE ARCHITECTURE EMERGES

Teams that reach this phase realize: the problem isn't the tool - it's the process. They build systems instead of prompting. Specs before generation. Evals on every output. Knowledge that compounds.

Spec-Driven Development Automated Verification Compound Learning

→ This is where the 4-Role Architecture begins.

The tool isn't broken. The process is.

And fixing the process requires more than better prompts - it requires defined roles, verified outputs, and knowledge that compounds.

Where is your team? Let's talk →

02 THE 4 ROLES
01

THE ARCHITECT

The Specifier

Deterministic Planning
"Ambiguity in the prompt is a bug in the code."

Decomposes vague requirements into rigorous, machine-readable specifications. Defines API schemas, data models, and architectural constraints before any code is generated.

ARTIFACT: spec.md - Single source of truth
02

THE BUILDER

The Generator

High-Velocity Implementation
"Code is a compilation artifact of the spec."

Executes spec.md using Agentic IDEs. Connects to internal tools to fetch live context. Produces code that strictly adheres to the project's style guide.

ARTIFACT: .cursorrules - Style enforcement
03

THE REVIEWER

The Auditor

Verification & Security
"Trust is good; Evals are better."

Static analysis against spec requirements. Dynamic evals with golden prompts. Blocks any PR that doesn't meet Context Compliance score.

ARTIFACT: eval.yaml - Regression suite
04

THE COMPOUNDER

The Optimizer

System Evolution
"A bug happens once. The second time, it's a process failure."

Root cause analysis on failures. Updates System Context to prevent recurrence. Manages the Corporate Brain - cumulative knowledge of decisions and patterns.

ARTIFACT: system_prompt.md - Living context
03 SERVICES
Agentic Transformation

AGENTIC TRANSFORMATION

Implementing the 4-Role Architecture in your engineering organization.

FIG.02
Architect 01

SPEC-DRIVEN DEVELOPMENT

Training Senior Engineers to write precise logic contracts. spec.md templates that guarantee high-fidelity agent output.

Builder 02

CONTEXT INFRASTRUCTURE

Your AI tools get secure access to your codebase, APIs, and documentation. No more context amnesia.

Reviewer 03

EVALOPS IMPLEMENTATION

CI for AI - automated evaluations on every agent-generated PR using LangSmith or custom pipelines.

Compounder 04

CONTEXT ENGINEERING

Monthly retainer: reviewing friction points, updating Agent System Prompts, making your entire dev team smarter over time.

04 CONTACT

Ready to build your Agentic Engineering Pipeline?

Let's discuss implementing the 4-Role Architecture in your organization.

kontakt@dezent.digital
Felix Schwenk
Felix Schwenk
Founder
Baden, Switzerland
Location
Baden, AG
Timezone
UTC+1
Status
Available
TECH GLOSSARY

For the technically curious:

MCP — Model Context Protocol
An open protocol that gives AI tools secure access to your internal systems - codebase, APIs, documentation. Developed by Anthropic.
spec.md
A machine-readable specification. Precise enough that an AI agent can execute it without hallucinating.
EvalOps
CI/CD for AI - the same automated quality gates you use for human code, applied to AI-generated output.
Agentic IDE
Development environments like Cursor or Windsurf where AI agents write code autonomously, not just autocomplete.
Context Engineering
The art of giving AI systems exactly the context they need - not too little (hallucinations), not too much (confusion).
System Prompt
The persistent instruction that defines how an AI agent behaves. Your "Corporate Brain" in text form.