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.