Ilya Lobkov
DODECAIDR

Hardware diagnostics and trade-in platform: native Apple clients, .NET 7–10 services and the Vue front end of the web platform. In the last six months — 11 repositories and five platforms.

▸macOS diagnostics client and secure distribution (Swift 6, AppKit): built the replacement for a legacy C# tool from scratch — 25+ hardware test modules (audio routing, ambient-light sensor, trackpad, keyboard and Touch Bar, Touch ID through IOReport, Bluetooth, camera, LCD, Neural Engine, Thunderbolt, fans, storage) on IOKit, SMC, IOReport and private HID APIs; universal binary, macOS 12+, localization parity across 9 languages. A privileged XPC daemon (SMAppService), TLS pinning, Keychain, Secure Enclave signed device-erasure certificates, Hardened Runtime and the full Developer ID pipeline with notarization; a bootable-USB wizard driving startosinstall erase-and-reinstall, gated behind a licence charge.

▸iOS and watchOS diagnostics (Objective-C + Swift, UIKit / SwiftUI): camera defect detection, battery, video and FeliCa modules driven entirely by backend configuration; white-label builds for partners, localization across 10+ locales, releases through fastlane + match. For Apple Watch — a full diagnostics suite in SwiftUI (MVVM): heart rate via HealthKit, motion via CoreMotion, speaker and microphone tests with FFT analysis (Accelerate / vDSP), Digital Crown, screen and charging; Watch apps drove a 30% increase in complementary software sales. Plus NSYSWatchTool — a Swift Package CLI for Watch fleet inventory with an MDM REST client and JSON / CSV reports.

▸Robotic test rig and macOS station: an app running about 60 automated tests on a robotic manipulator — cameras, TrueDepth and LiDAR (ARKit), Face ID, NFC, sensors, wireless charging instead of USB, speakers without clipping. The robot protocol over RabbitMQ / AMQP: a four-step handshake, typed outcomes, structured logging to Sentry / GlitchTip; automatic Wi-Fi provisioning (NEHotspotConfiguration) cut connection losses from 80% to about 5%. The diagnostic station is a menu-bar AppKit app (Tuist) controlling iOS and Android devices through the private MobileDevice.framework, libimobiledevice / usbmuxd and adb: genuine-parts verification, battery health, DFU / recovery restore, the Mac ↔ iPhone TCP protocol, VPP licences and iOS 26 compatibility.

▸Full-stack platform work (ASP.NET Core 7, EF Core, SQL Server, Vue 2 and Vue 3): new licence types and charging endpoints for laptop and headphone diagnostics, test-result models from EF entity through controller and view model to Vue component, a History Headphones section with filters and contract tests, file-upload limits across app and nginx. Microservices: a new .NET 10 service receiving headphone-diagnostics results (layered API / BL / DAL, health checks, ProblemDetails, xUnit) with an upsert race condition eliminated; maintenance of .NET 9 services — device registration with resale-blocker flags, EF migrations, secrets moved to environment configuration, and a fixed race in the access-code pool.

▸Network monitor and internal tooling: a cross-platform monitor in Python + PySide6 and Vue 3 pings company services and reference hosts every second, attributes failures to the local network, the ISP or company servers using a sliding loss window, measures throughput, journals to SQLite and renders a PDF report with charts in 6 languages; packaged with PyInstaller for Windows and macOS. Plus a security-hardened internal admin dashboard (Python backend + JavaScript SPA) with CSRF tokens, strict CSP with nonces, Origin / Host validation, secret masking and sandboxed subprocess execution.

▸Performance, architecture and CI/CD: cut CPU load of the flagship macOS app from 45% to 4% through memory-leak resolution and refactoring; led the migration to Clean Architecture, cutting cross-team feature delivery from a month to about 1.5 weeks. Fastlane and GitLab CI: automated IPA / DMG builds, match code signing, Sentry, auto-changelogs to Telegram, plus documentation and merge-request standards for the team.

▸Apple enterprise and team: I administer MDM, Apple Business Manager and VPP for a global organization, and wrote Python tooling for Apple's VPP API plus a 46-endpoint multi-region Apple service status monitor. I champion Claude Code, Cursor and OpenAI Codex for large refactors, test generation and internal tooling; I mentor developers and run code reviews, one-on-ones and retrospectives across the iOS and Android teams.

My own product studio, run alongside a full-time role. Solo full cycle from idea to release: MeteoHealth (iOS + watchOS, HealthKit and an on-device statistical engine), Day of the Day (macOS menu bar), ProxAlert (location-based alerts), FateBall. Next.js/TypeScript websites — dodecaidr.pro, meteohealth.pro, runcalculator.pro and Hookah Hub (14,000+ entries, 5 languages) — self-hosted on Docker. Currently building Lanternly: a privacy-first journal for iPhone, iPad, Mac and Apple Watch on Swift 6, SwiftData and Apple Foundation Models.

Solo, full cycle: MeteoHealth, Day of the Day, ProxAlert and FateBall shipped to the App Store; dodecaidr.pro, meteohealth.pro, runcalculator.pro and Hookah Hub live in production. Product, code, release, ASO, localization and marketing — single-handedly.

Public transport app for several cities. Cut memory usage by 60% — from 650 MB down to 260 MB. Shipped Bluetooth beacon (iBeacon) payments, improving transaction reliability by 20%. Set up CI/CD with Fastlane: the build-and-release cycle went from a full day to 2 hours. Led a UI redesign with designers under regulatory and accessibility requirements, and ran architecture reviews around SOLID and VIP.

App for buying, renting and maintaining residential property. Rebuilt the chat architecture from scratch — 100% QA acceptance and zero chat-related defects. Shipped a property viewing, purchase and rental module that opened a new revenue stream, a tenant self-service module (lease renewal, utility payments) and flexible authorization with guest and authenticated states.

Digital menu application: ordering, payment processing, real-time order status tracking and personalized user accounts.

Client projects delivered end to end: an educational Apple TV (tvOS) app and an e-commerce app for an online clothing store. Stack: UIKit, YouTube API, MVVM, URLSession, Codable.

GPA 4.72 out of 5. Activities and societies: the Sporadic radio club, the Olimp fight club, the Kazachki choir. Hobbies: whitewater touring, climbing, mixed martial arts, programming, research, writing short stories; graduated from music school (accordion). Also: 4 academic publications, 3 research projects in progress, participation in annual radio expeditions.