Architecture and technical direction
Own the end-to-end architecture of the product: data model, services, integrations, infrastructure, security, and reliability. Communicate architecture through C4 diagrams (context, container, component, and code where it adds value) that stay current with the system. Author and curate ADRs (architectural decision records) that capture context, options, trade-offs, and consequences for every load-bearing decision. Set the architectural guardrails — boundaries, contracts, SLAs — that let the team move fast without re-litigating foundational decisions. Make build/buy/integrate decisions across platforms, frameworks, third-party services, and infrastructure. Define and evolve the canonical schema, versioning model, and core abstractions of the product
Write production code on critical paths, especially where the architectural risk is highest. Lead design reviews and review PRs on load-bearing parts of the system. Prototype hard problems quickly to de-risk decisions before the team commits. Profile, debug, and stabilize systems under real production conditions
Reliability, security, and enterprise readiness
Drive the architecture for deployment, data residency, and audit. Lead the response to systemic reliability issues and convert them into durable architectural improvements. Define the access control, permissions, and audit-log model across the product. Own incident postmortems on systemic issues and translate them into roadmap-level changes
Cross-functional and engineering collaboration
Partner with engineering leadership to translate architecture into shippable increments. Work directly with Product on what’s architecturally feasible inside the current quarter vs. what’s a quarter-out investment. Mentor engineers on systems thinking, trade-off analysis, and writing for design clarity. Represent the product’s architecture to enterprise customers during technical evaluations and security reviews
AI-augmented engineering practice
Use modern AI engineering tooling (IDE agents, CLI assistants, MCP servers) as a daily working layer. Set the team’s working patterns for spec-driven development, evaluation gates, and AI-assisted code review. Distinguish where AI accelerates the work and where it introduces risk; encode that distinction in the delivery process.