The mandate: Build the core systems behind millions of AI-driven conversations — real-time audio, global telephony, prompt orchestration, testing frameworks, observability, and the developer-facing API surface. You work directly with the founding team and you own projects from architecture through production. The honest framing: this is a backend infrastructure job at an AI company, not an AI job. No ML required. The value proposition to a strong backend engineer is the systems problem — real-time, concurrent, latency-bound, at genuine scale — plus a small team and founding-adjacent ownership. One thing the JD includes that's unusual for a backend req and worth surfacing early: customer onboarding support and troubleshooting are named responsibilities. Not the majority of the job, but a pure-heads-down engineer who wants zero external contact will find that friction. Screen for willingness. What You'll Own
-
Foundational platform architecture: prompt orchestration, testing frameworks, observability dashboards, telephony integrations
-
Scalable backend systems and DevOps workflows that hold across millions of calls
-
Complex projects end to end — architecture through production, including delivery and monitoring
-
Intuitive APIs and interfaces built with product and design for fast customer onboarding
-
Customer onboarding support, troubleshooting, and converting feedback into product improvements Requirements Hard gates:
-
2+ years backend development with strong CS fundamentals (CS/CE/EE degree or equivalent)
-
Real experience with scalable, high-performance backend systems — concurrency, data, storage, or distributed components
-
On-site in Redwood City
-
Can code and design live. Two 45-minute technical rounds plus a three-round onsite. Profile:
-
4–8 years total for the senior bar (the JD's 2+ minimum is the floor, not the target — the comp band signals senior)
-
Founder-style ownership. The JD says "take ownership like a founder."
-
Thrives in ambiguity and learns new problem domains fast
-
Communicates clearly and drives work forward across teams Strong bonuses:
-
Real-time or streaming systems — WebRTC, WebSockets, SIP, media pipelines. Closest match to the actual problem.
-
Telephony — Twilio, Vonage, Telnyx, SignalWire, carrier integrations, SIP trunking
-
Infrastructure, database, or large-scale backend systems work (explicitly called out in the JD)
-
Distributed systems depth: queues, event streaming, consensus, sharding
-
Developer tooling and API design for external developers
-
Observability and DevOps: tracing, metrics, incident response
-
Startup experience, especially early-stage Location and visa:
-
On-site, Redwood City
-
Sponsorship: Yes — H-1B, TN, L-1, E-3, F-1 (OPT/CPT), and O-1 Anti-patterns
-
CRUD-only engineers with no concurrency, latency, or distributed systems exposure
-
Big-company engineers who've only worked inside a mature platform with a dedicated infra team
-
Engineers who need a spec, a ticket, and a defined scope
-
Anyone who wants zero customer contact — onboarding support is in the JD
-
Slow-cycle enterprise engineers on quarterly release trains
-
Frontend-primary engineers. The loop is explicitly backend and technology-agnostic; React specialists go to a different req.
-
Remote requirements Who Will Thrive Here
Someone who's debugged a race condition in production at 2am and enjoyed it. They care about p99 latency, they've built something that had to be correct under concurrency, and they want to own a system rather than a service someone else designed. Early-stage startup engineers and infrastructure people from companies where reliability was the product fit naturally. The real-time constraint is what separates the people who'll love this from the people who'll be bored.