Get a reply from this employer
Your resume and cover letter, tailored to exactly what they’re hiring for.
Zeo Consulting is seeking an architecture-focused software leader to design system boundaries, data models, and contracts that work across brand experiences while keeping determinism where needed. The role emphasizes agent-driven development and rigorous verification before merging, with a strong emphasis on typed schemas and observability.
Ideal candidates will drive architecture decisions in collaboration with product and engineering, and will craft specifications, RFCs, and invariants to
This is an architecture role that writes code, not a coding role that occasionally diagrams.
The hard problems here are not framework problems. There are questions like
What is the right abstraction for behavior that must stay identical across brands with wildly different experiences?
Where does a system have to be deterministic and reproducible, and where can human judgment live?
What belongs in a shared contract versus a local implementation?
How do you make a system's invariants checkable rather than aspirational?
We need someone who thinks in systems and concepts first, and who can then build the thing. Specifically
Define service boundaries, data models, and contracts between components
Design seams that let the same code run locally and hosted without a rewrite
Turn fuzzy capabilities into clear models of core entities and rules
Apply MIT concept design methodologies to agentic system design
Design abstractions that keep behavior consistent across brands with very different experiences
Decide where systems must be deterministic and where human judgment can apply
Make architecture decisions alongside founders with product and engineering backgrounds
Plan up front how each system will prove its correctness, rather than retrofitting tests
Make system invariants checkable through code and tests
Write specifications, decision records, and product invariants in RFC 2119 language
Direct AI agents to build features and review their output in depth before merging
Write specs that agents can execute and reviewers can check
Maintain agent instructions and skills as first-class artifacts
Run independent adversarial reviews before opening pull requests
Catch plausible but wrong results (e.g. silent safety gates, empty version pins, hollow dependency directories)