Hiring: Mobile Engineer (React Native / Expo) - Real-Time Voice AI

The main focus will be around building an iOS app first and an android app 2nd.
Mobile Engineer (React Native / Expo) - Real-Time Voice
Contract · Fully remote · Vietnam-based · Working with a US Eastern founder
About us
ObMob helps physicians prepare for medical board exams. Our flagship feature is a live voice examiner - a spoken mock oral exam with an AI that listens, responds, and interrupts naturally. Live and revenue-generating.
Stack: React/[protected info], Supabase, Pipecat voice pipeline (Deepgram → Claude → Cartesia).
You'd be our first dedicated mobile engineer, working directly with the founder.
The role
We have a working web product. We need real iOS and Android apps where the live voice examiner feels as good as it does in a browser. That's not a port - it's a rebuild against native audio sessions, OS interruptions, Bluetooth routing, and background execution rules that have no browser equivalent.
Split: -65% mobile voice · -20% monetization and releases · -15% DevOps.
What you'll do
Mobile voice (-65%)
Build and own our RN/Expo apps, migrating off Capacitor (we need native audio control, not a WebView)
Evaluate WebRTC (Daily) vs telephony/SIP (Twilio/Telnyx → Pipecat) on real devices, recommend with numbers, then ship it
Make voice rock-solid on real phones: CallKit / ConnectionService, background and screen-off continuity, Bluetooth/wired/speaker routing, echo cancellation, barge-in and end-of-turn detection
Handle the ugly cases: incoming calls mid-session, Wi-Fi→cellular handoff, OEM battery managers, long-session battery and thermals
Shared layers - TypeScript, Supabase/PostgreSQL, auth, deep links, push
Monetization and releases (-20%)
Sessions are paid for with credits, and we want to keep those off in-app purchase where we can - store commission on an AI product with real inference costs is the difference between a healthy margin and a bad one. You'd own:
A server-side credit ledger as the single source of truth; the app spends entitlements, never mints them. Web purchases reflect in the app immediately.
Paywall UI that's region-gated and remotely configurable, so store policy changes don't require a resubmission
App Store / Play submissions, signing, releases, reviewer accounts, and the appeal if we get rejected under 3.1.x
Privacy declarations: mic permission, App Privacy labels, Play Data Safety and Health Apps declaration
DevOps (-15%)
CI/CD and EAS Build/Submit/Update, monitoring and backups, Sentry, voice telemetry (latency percentiles, disconnect rates), secrets and access hygiene.
Required
4-6 years professional development; strong TypeScript and React
Shipped production apps to the App Store and/or Play, and ran the submissions yourself. RN/Expo preferred.
Hands-on real-time audio is the #1 requirement - production WebRTC and/or native audio-session work (CallKit, ConnectionService, background audio, Bluetooth). Daily, LiveKit, Twilio, Agora, Pipecat all count.
Comfortable dropping into native code when a JS library falls short
Supabase or Firebase, and SQL
Working English, including enough written precision to handle an App Review appeal
Strong async self-direction
Nice to have
App-store billing policy in practice (link-out flows, entitlement sync, a rejection you got overturned) · Telephony/VoIP/SIP · Pipecat or audio DSP · [protected info] and monorepos · Health-tech or AI voice interest
Terms
Contractor, invoiced monthly. Initial term [6 months], expected to extend.
[Full-time / -30 hrs/week] · [X,XXX-X,XXX USD/month] depending on audio depth
First store submission target: [X] months from start
Test-device budget reimbursed (modern iPhone, recent Android flagship, budget Android, Bluetooth headset)
Developer accounts stay under our entity; you're added by role. Standard contractor agreement with IP assignment. Paid via [Wise/Deel], fees covered.
How we work
Fully remote and async-first. We're US Eastern; Vietnam is -11-12 hours ahead. We ask for a few hours of weekday overlap (your evening, our morning) - the rest is flexible. Direct founder access, fast feedback, work that ships.
We're health-tech, so we set you up properly from day one: least-privilege credentials, staging for daily work, MFA, production access per task. You will never be asked to handle patient data.
To apply
Short intro and why this fits
Portfolio, GitHub, or shipped apps - especially voice, calling, or audio A hard audio or real-time bug you solved - symptom, how you isolated it, actual cause. This is what we read most closely.