Full Stack Engineer

Location: Hybrid- NYC (3+ days in office required) Type: Full-time W-2 employee, subject to a 90-day introductory period Salary and equity commensurate with experience and performance --- About AviaryAI We build AI voice agents for banks, credit unions, and insurers. Our agents are live and handling real customer calls today. We’re backed by Y Combinator and growing. We started with voice. Our SMS and RCS agents began as clones of the voice product — and they’re growing fast. Messaging is becoming a major part of our platform, and it needs a dedicated owner --- The Role You’ll be a full stack engineer working across our platform — and you’ll head the engineering effort on our SMS and RCS expansion. This means: Leading the engineering effort to evolve our SMS and RCS agents from voice-agent clones into a messaging-native product — rich cards, suggested replies, media, session handling Writing the backend logic that manages conversation state and decision-making (Python) Building internal dashboards and tools to configure and debug agents (TypeScript) Contributing to our core frontend application as needed Integrating with messaging providers and carrier infrastructure — delivery, 10DLC, opt-in/opt-out Debugging real production issues when things break You’ll own the messaging vertical from prototype to high-volume production. This is not a ticket-driven role. You’ll have a lot of responsibility and a lot of autonomy. --- What We’re Looking For 1–3 years full stack experience (new grads welcome) Strong in Python Comfortable in TypeScript / modern frontend frameworks Experience building systems with real state or multi-step workflows Comfortable integrating third-party APIs Able to work without detailed specs Bonus if you’ve worked in: Startups Fintech / insurance Messaging or telephony systems (Twilio, Telnyx, Bandwidth, SMS compliance) LLM applications (prompting, agent frameworks, evals) --- What This Is Like You won’t just be building CRUD endpoints. You’ll be building systems that: Decide what to say next in a live conversation — hours or days after the last message Enforce business rules and compliance requirements (TCPA, carrier filtering, quiet hours) Handle messy, real-world edge cases Run thousands of concurrent conversations reliably Your code will power real conversations with real customers — and unlike voice, every word is in writing.