Software Engineer

Avoice is building the AI-native workspace for architecture and engineering firms.

Backed by leading investors and Y Combinator, we're rethinking how one of the world's largest industries operates. Architects and engineers spend the majority of their time not designing, but handling documentation, coordination, QA/QC, and compliance, workflows that are still largely manual.

The Role

We're looking for engineers who combine strong fundamentals with an AI-native mindset. You should be someone who can build robust, production-grade systems — and also understand how to rethink those systems in a world where intelligence is embedded everywhere. You will:

• Build AI-native features end-to-end (backend, systems, and product)
• Design agent-based workflows (reasoning, tool use, orchestration)
• Work closely with founders on architecture and product direction
• Ship quickly and iterate with real users
• Turn complex, messy workflows into scalable systems

What We're Looking For

Strong engineering fundamentals

• Proven ability to build and ship production systems
• Strong understanding of backend systems, APIs, and data models
• Ability to write clean, maintainable, and scalable code
• Solid grounding in system design and tradeoffs (latency, reliability, cost)
• Comfort debugging real-world issues in production

AI-native mindset

• Experience building with LLMs beyond simple wrappers
• Ability to decide what should be handled by models vs deterministic systems
• Understanding of evaluation, failure modes, and reliability in AI systems

Product sense

• Ability to translate ambiguous workflows into usable products
• Strong intuition for UX and workflow design
• Comfortable operating in fast-moving environments

Nice to have

• Experience with BIM (e.g. Revit, Archicad)
• Experience with agent systems or orchestration
• Experience with document-heavy or unstructured data workflows
• Startup experience

Why This Role

• Large, underserved industry with real demand
• Opportunity to define how AI-native software is built from first principles
• Deep, complex problem space — not a thin wrapper around existing tools
• Early team — high ownership and direct impact