Illya Lukinov

Lead Software Engineer · AI Adoption & Developer Productivity

I ship software, and I help teams ship faster with AI. 15+ years across back-end and quality engineering — now focused on AI adoption, agent tooling, and the discipline that makes AI-augmented development actually stick.


01

AI Enablement

On my team I'm the internal driver of AI adoption — teaching disciplined AI use (spec-driven development, Definition-of-Done gating for AI output, agent-assisted code review), building custom agent skills and CLAUDE.md / AGENTS.md repo scaffolding so agents plug into real stacks, and using the same workflows to ship the projects below.

I care about AI output that holds up in production, not just demo-friendly code. The short version: help teams skip the enthusiastic-but-chaotic phase and get to shipping more, with guardrails.


02

Lab

I ship using the same AI-augmented workflows I help teams adopt. These are recent experiments — each one a way to push on a different slice of today's AI toolkit: on-device models, multimodal live APIs, agent-augmented UI.

EchoAct

On-device voice dictation for macOS. Menu bar app with Parakeet TDT v3 on the Neural Engine, custom LLM actions, and on-device Apple Translation. No audio ever leaves your Mac.

SwiftmacOSNeural EngineOpen Source
VibeDeck

A real-time "vibe coding" environment. Describe a vision via voice using the Gemini Multimodal Live API, watch it come to life, and navigate the whole thing with a gamepad.

ReactGemini MultimodalVoice AIGamepad
Interactive AI CV

A video-first resume with a Gemini-powered AI agent that answers questions about the candidate. Timeline syncs to relevant experience sections as you watch the pitch.

ReactGemini APIFirebaseVideo
ChromaFlow

Continuous, distraction-free time tracking with a living canvas, planned-vs-actual flow, and timestamped session notes pinned directly to the timeline.

ReactTypeScriptFirebaseFramer Motion

03

About

I'm a Lead Software Engineer based in Salt Lake City. I came up through back-end and quality engineering — which is why I care about AI output that actually holds up in production, not just demo-friendly code. I teach from inside the SDLC, not around it.

Outside of software I design and build custom ergonomic keyboards from scratch: PCB design, 3D-printed cases, soldering, firmware. I guess I like building things end to end even when nobody's paying me.

AI-Assisted Dev Claude Code, GitHub Copilot, Cursor, Google Antigravity, Gemini API
AI Integration Custom agent skills, CLAUDE.md / AGENTS.md scaffolding, rapid UI prototyping
Backend & Cloud Java, Spring Boot, Python, Kubernetes, AWS, PostgreSQL
Interests AI evangelism, spec-driven development, developer tooling, ergonomic keyboards

04

Get in Touch

I'm open to AI enablement contracts, productivity-engineering roles, and conversations with teams trying to get more out of AI-augmented development. Feel free to reach out.

Say hello →