Israr K.

Need voice AI inside a mobile app, or a React Native app that actually holds up in production? That's what I build. I've shipped apps used by millions: UPI payments at Flipkart (10M+ users), and I led the native rebuild of Wadhwani Skilling (200K MAU, 5 countries), including its AI voice coach for pitch and interview practice. My most recent Upwork contract was a custom wake word and voice biometrics integration in a mobile app (5 stars). 𝗖𝗹𝗶𝗲𝗻𝘁𝘀 𝘂𝘀𝘂𝗮𝗹𝗹𝘆 𝗰𝗼𝗺𝗲 𝘁𝗼 𝗺𝗲 𝘄𝗵𝗲𝗻 - They want real-time voice AI in their app: talking to an AI, live speech-to-text, wake word, voice commands - Their app is half-built, slow or crashing, or the last developer left - It was built with AI tools (Cursor, Lovable, Bolt) and breaks once real users arrive - They need the backend and AWS done too, not just the screens 𝗩𝗼𝗶𝗰𝗲 𝗔𝗜 𝗜'𝘃𝗲 𝘀𝗵𝗶𝗽𝗽𝗲𝗱 - Wadhwani Skilling: AI Pitch Trainer and Interview Coach with Azure Speech, used across 5 countries. Cut pitch video uploads from ~25MB to under 5MB per minute so it works on low-end phones. - Relyve: a voice companion app for early memory loss that I designed and built on my own for the RevenueCat Shipaton, now live. Real-time voice with Gemini Live over WebSockets, with tool calls that write to the database, interruption handling, scheduled reminders and RevenueCat subscriptions. React Native, Node.js and MySQL on AWS. - rn-whisper-stt: my open-source React Native library for on-device, offline speech-to-text with whisper.cpp. Live partial results, about 100-400 ms per phrase on mid-range phones, no per-minute API cost. Voice stack: OpenAI Realtime API, Gemini Live, LiveKit, Twilio, Azure Speech, ElevenLabs, Whisper (on-device), wake word detection, WebSockets and WebRTC. 𝗖𝗼𝗻𝗻𝗲𝗰𝘁𝗲𝗱 𝗱𝗲𝘃𝗶𝗰𝗲𝘀 - Fodxpert: a digital signage platform running 500+ restaurant screens across India. I built the Android TV player (React Native for TV with Kotlin native modules: auto-start on boot, offline video playback, remote logs and diagnostics), a Raspberry Pi player, the mobile CMS, the web dashboard and the Node.js API, with MQTT for real-time updates, PostgreSQL on RDS, EC2, and CloudWatch and Grafana alerts. - BLE apps with native modules on iOS and Android, including an enterprise Bluetooth integration for Honeywell. 𝗪𝗵𝗮𝘁 𝗜 𝗰𝗮𝗻 𝗱𝗼 𝗳𝗼𝗿 𝘆𝗼𝘂 Mobile - React Native and Expo apps from scratch, or take over an existing one and stabilize it - Performance: I rebuilt a 200K-MAU app from WebView to native, cutting screen loads from 5.5s to 1.2s and data per screen from 3.9MB to 6KB. It runs at 99.83% crash-free sessions - Native modules in Kotlin and Swift - Push notifications, deep links, payments and subscriptions (RevenueCat, in-app purchases, Stripe) - Builds and releases with EAS, Fastlane and TestFlight, OTA updates, and App Store / Play Store submission AI - Voice agents and real-time conversation in mobile apps - AI features that call your backend and database through tool calls - On-device speech-to-text when privacy or API cost matters Backend and infrastructure - REST and GraphQL APIs in Node.js and Express - PostgreSQL, MySQL, Supabase, Firebase - AWS (EC2, RDS, Secrets Manager, CloudWatch), Docker, CI/CD with GitHub Actions Production health - Sentry, Crashlytics, CloudWatch and Grafana, p95 metrics, phone alerts - Jest, React Testing Library, Maestro and Detox Industries: fintech and payments, edtech, health and wellness, media and streaming, B2B SaaS, retail and digital signage. Also worked on: HDFC PayZapp onboarding and referrals, STC Play (1M MAU), and open-source contributions to iNaturalist. 𝗛𝗼𝘄 𝗜 𝘄𝗼𝗿𝗸 Daily updates, test builds you can install on your own phone, clean code in your repo from day one, and I don't disappear mid-project. Send me a short note on your app (what it does, what's broken or what you want to add) and I'll reply with how I'd approach it and a rough timeline.