I will review your AI automation, API-heavy workflow, or multi-step backend process from a systems and reliability perspective — not just whether the happy path works.

I examine workflow state, integrations, failure modes, retries, duplicate execution, idempotency, AI vs deterministic logic, approval boundaries, model/API cost exposure, sensitive-data boundaries, and operational clarity.

You receive a concise architecture/reliability review with prioritized findings, practical recommendations, and a target workflow/architecture direction where useful.

This service is an architecture and reliability review. Production implementation is not included in the fixed price. Please do not send passwords, API keys, tokens, private keys, or other secrets. Sanitized documentation, diagrams, logs, and code excerpts are preferred.

Deliverables

Workflow & architecture map — current-state system boundaries and key state transitions.

Reliability findings — ranked failure, replay, integration, AI-boundary, cost, and operational risks.

Prioritized action plan — what to fix now, what to defer, and why.

Target architecture direction — a safer/simpler design where the existing structure needs material change.

Client requirements Required: describe the workflow and business outcome; describe the main failure, risk, or problem; provide sanitized documentation, diagrams, or code excerpts. Optional: sanitized logs, API list, model cost concerns, and duplicate/retry issues.