Scoped work · Clear outputs
How to engage me.
Independent senior architecture and engineering leadership, scoped to the decision in front of you rather than another full-time hire.
Each engagement below has a defined duration, a defined scope, and a defined output, so you know what you get before you start.
Four scoped ways to start · duration and outputs on each entry
Browse engagements-
Architecture Decision Sprint
For a consequential architecture decision where the organization needs an independent technical position.
Typical situations: cloud or platform selection, modernization direction, build-vs-buy, distributed architecture, integration strategy, or an unresolved architecture disagreement between teams.
- Architecture assessment
- Options & trade-offs
- Risk register
- Recommendation
- ADR
- Executive readout
Example deliverable — ADR excerpt
ADR-014 — Adopt event-driven integration over point-to-point sync Status: Accepted Context: Three downstream systems each poll the core ledger on a fixed schedule; polling load is now the ledger's top scaling risk. Decision: A single Kafka topic becomes the system of record for balance-change events; downstream systems consume, not poll. Consequences: Downstream teams own their own consumers; the ledger no longer needs to know who is listening.
-
Architecture & Production Readiness Review
For a system approaching production, or one already experiencing scaling or reliability problems.
Reviews architecture, security, data, resilience, observability, performance, deployment, and operational readiness across the system, then delivers findings ranked by severity with supporting evidence.
- Findings
- Severity ranking
- Evidence
- Remediation roadmap
Example deliverable — findings register entry
Finding F-07 — Sev 2 (Resilience) Observation: Checkout calls the payment gateway with no timeout budget or circuit breaker. Risk: A slow gateway response holds a request thread until the platform-wide timeout, degrading unrelated traffic. Remediation: Add a bounded timeout and circuit breaker; fail fast to a queued-retry path instead of blocking the request.
-
AI Production Readiness
For an organization that has built an AI or agent proof of concept and now needs to answer: can we safely put this in production?
Assesses evaluation, authorization, identity, tool access, data boundaries, human approval, observability, failure behaviour, and auditability against what production actually demands.
- Evaluation
- Authorization & identity
- Tool & data boundaries
- Human approval design
- Observability
- Auditability
Example deliverable — approval gate spec
Gate: Refund-issuing action Trigger: Agent proposes a refund over $500, or to a new payment method not on file. Requirement: Human approval before execution; decision and rationale logged to the audit trail. Covered failure mode: Agent proposes an out-of-policy refund after a prompt-injection attempt — gate blocks execution regardless of the agent's stated justification.
-
Strategic Architecture Advisory
For CTOs and VPs who need independent senior architecture support without adding another full-time executive.
Ongoing coverage across architecture reviews, technology decisions, modernization roadmap, design escalation, engineering risk, and vendor evaluation, brought in as the questions arise rather than on a fixed deliverable schedule.
- Architecture reviews
- Technology decisions
- Modernization roadmap
- Design escalation
- Vendor evaluation
Example deliverable — monthly review, one page
This month: 2 design escalations resolved, 1 vendor evaluation in progress (event-streaming platform), 1 open risk carried forward (data-residency question on the EU rollout). Standing channel: architecture-escalations — median response time under 4 business hours.