piroznoe

Official Polkadot network mobile app, commissioned by Parity Technologies: self-custodial identity with decentralized human verification (Proof of Personhood). • Engineered a multi-party WebRTC video attestation system: hand-written Metal shaders for real-time video rendering, gesture-driven verification UX, and frame-accurate animation sync across participants; traced and eliminated rare concurrency races in the media pipeline. • Implemented the decentralized messaging core over Polkadot Statement Store — end-to-end encrypted, per-peer, multi-device, with no chat server — and designed the delivery layer that makes it reliable: an inbox/outbox state machine with retry classification, debounced batching, and resend on reconnect. • Owned chat UI architecture and performance, building it into the platform other engineers shipped features on — in-chat extensions, calls, and Proof of Personhood flows all run on its diffable data source layer. On- device search uses SQLite FTS5, chosen after benchmarking three approaches on 100K messages: 0.2– 2.2 ms against 59–64 ms for Core Data predicates and 223–245 ms for regex. • Contributed a resilient WebSocket transport to the open-source Substrate iOS SDK, holding connections through LTE/Wi-Fi transitions via viability checks, pong timeouts, coordinated connection state, and deferred reconnects. • Coordinated and shipped iOS beta releases ahead of Web3 Summit, contributing to a daily release cadence sustained for six months across a 100-person test cohort at 99.7% crash-free sessions, with zero crashes across ~1,000 conference users; led the CocoaPods to SPM migration. Tech Stack: Swift, UIKit, SwiftUI, WebRTC, Metal, Substrate SDK, Polkadot Statement Store, SCALE, Core Data, SQLite FTS5, Swift Concurrency, SPM, Firebase Links: Polkadot App (polkadot.com/app) · substrate-sdk-ios (open source)

Simble: on-demand vehicle insurance app that automatically starts insurance when driving. KIWU: UK-focused trip tracking app providing driving statistics and analytics. • Led a cross-functional team of 6 developers (iOS, Android, backend, web) across two production apps; hired, onboarded, and mentored 3 iOS engineers, progressively delegating feature and code ownership as their code quality matured, while owning client-facing technical delivery for four years. • Architected a shared iOS framework for trip tracking, sensor processing, and networking used by both products, eliminating duplicated platform work so a single fix shipped to both apps. • Modularized shared UI and feature code into independently buildable Swift packages with an isolated SwiftUI-preview workflow for UIKit — incremental UI build time fell from 30 s to 5 s across both products, laying the groundwork for automated visual regression testing. • Partnered with analytics and ML engineers to diagnose and improve driving-behavior classification from Core Location and Core Motion sensor data. Tech Stack: Swift 5, UIKit, SwiftUI Previews, SPM, CoreLocation, CoreMotion, CoreData, Alamofire, Protobuf

Mobile banking app with payment processing, money transfers, and budget tracking. • Designed and built a modular home screen mimicking iOS SpringBoard: folders, drag-and-drop, and custom shortcuts. • Created advanced UI animations and seamless transitions with CoreAnimation. • Designed a dynamic, adaptive collection view layout for users with visual impairments: enhanced readability, larger touch targets, and simplified navigation (accessibility-first design). Tech Stack: Swift 5, UIKit, CoreAnimation, MapKit, Alamofire, APNS

Food ordering and restaurant reservation app with multilingual support (4+ languages). • Developed a WebSocket interface for real-time order tracking and live status updates. • Optimized networking and payload handling so the app stays responsive on low-bandwidth (EDGE) connections. Tech Stack: Swift 4, UIKit, Starscream, CoreAnimation, MapKit, CoreLocation

Self-directed product: an augmented-reality chemistry learning app for school students, built on ARKit from its first public beta. • Adopted ARKit from the WWDC 2017 beta, working around early tracking and lighting-estimation limitations to keep virtual models stable in a classroom environment. • Owned the full product cycle solo: requirements from chemistry teachers, iterative testing with school students, App Store preparation. Tech Stack: Swift 4, ARKit, SceneKit, UIKit, CoreAnimation