antonnovoselov

• Shipped open source iOS speech-to-text app on App Store, featuring system-wide AI voice keyboard for any app. • Integrated 20+ AI providers with strong on-device support (LLM: Apple Foundation Models, MLX; ASR: WhisperKit, NVIDIA Parakeet); cloud LLMs (OpenAI, Anthropic, Gemini, Mistral, Groq); cloud ASR (Deepgram, ElevenLabs, Soniox, Cohere); OAuth (ChatGPT, Gemini, GitHub Copilot). • Architected on-device RAG pipeline for Smart Search - semantic search across all notes by meaning, with embedding and retrieval fully on-device. • Implemented chat-with-your-notes (single-note, multi-note, Smart Search modes) with LLM tool calling (cross-note search, web search, reminder extraction) - powered by Apple Foundation Models on-device or cloud AI. • Engineered AI processing engine with 40+ presets and per-preset sampling configuration across cloud and on-device providers. • Designed multi-target architecture (6 targets: main app, keyboard, widgets, watch, share, action extensions) with cross-process IPC via Darwin Notifications. • Integrated App Intents framework: Siri/Shortcuts, chainable App Entities, Spotlight indexing, widget and control actions. • Built cross-device sync: WatchConnectivity for watchOS companion; SwiftData/CloudKit across iPhone, iPad, Mac.

• Migrated paywalls from WebView to native SwiftUI, boosting performance and increasing paid conversions by 5% on a user base of 10 million. • Worked through the iOS codebase's full modernization: 100% UIKit at start (2022) to fully SwiftUI by end (2025), with concurrency migrating from GCD to Swift Concurrency. • Delivered key features including Offline Mode, In-App Purchases, Push Notifications, and custom Sounds & Haptics modules. • Migrated iOS networking from REST to GraphQL, eliminating constant back-and-forth with backend engineers for endpoint changes and accelerating feature delivery across the iOS team. • Refactored codebase from MVC/MVVM to TCA architecture, increasing modularity and accelerating development velocity. • Expanded unit and UI test coverage across critical user flows, improving reliability and confidence in releases. • Streamlined release process via Bitrise CI/CD and GitHub Actions, reducing manual overhead. • Improved ASO, resulting in higher search-driven organic traffic and greater app visibility in the App Store. • Drove Git Flow adoption across the iOS team, smoothing hotfix handling and release cadence.

• Built and shipped new app features in Swift, SwiftUI, UIKit, and VIPER. • Rebuilt the analytics module with batched event sending, significantly reducing network requests and improving app performance. • Refactored a legacy Objective-C codebase into Swift. • Introduced unit tests to a previously untested codebase, establishing a testing foundation.

• Built iOS app to control video conferences hosted on Cisco Acano Meeting servers using Swift, UIKit, and WebSocket networking. • Refactored legacy code and resolved bugs to enhance app stability, performance, and maintainability.