Dor Luzgarten
Mobile Developer 🔄 AI Engineer
I've been building mobile apps for a while - iOS, Android, Flutter. Along the way I also built the backends behind them - APIs, databases, distributed systems, the stuff that has to actually hold up under load. At some point I got more interested in AI, and now most of my work is around agents and automations. I like building things that actually do stuff on their own.
Experience
Mobile Development
2018 – 2026
Built mobile apps on iOS, Android, and Flutter - both native and cross-platform. Worked on consumer apps with real scale, handled architecture decisions, and dealt with App Store and Play Store pipelines.
- Native iOS with Swift & UIKit / SwiftUI
- Native Android with Kotlin & Jetpack
- Cross-platform with Flutter & Dart
- CI/CD with Fastlane, GitHub Actions, Bitrise
Backend Development
2020 – 2026
Built and maintained backend services for mobile apps. REST and GraphQL APIs, databases, third-party integrations. Worked on everything from simple serverless functions to containerized services.
- RESTful & GraphQL API design
- PostgreSQL, Firebase, Supabase
- Node.js / NestJS, Python / FastAPI, Go / Gin
- Docker, AWS Lambda, GCP Cloud Run
AI Agents
2024 → Present
Building regular apps started feeling too predictable. Agents actually do things - reason, use tools, make decisions. Coming from mobile means I can ship agents that live inside the apps people already use, not just in a terminal.
- Multi-agent orchestration with LangGraph
- RAG pipelines and vector stores
- Tool use and function calling with OpenAI & Anthropic APIs
- Autonomous research, coding, and data extraction agents
- GenUI in Flutter - UI that renders itself based on model output
- Apple Foundation Models Framework - on-device LLMs, no server needed
OpenClaw
2026
My own agent setup running on a private VPS. Everything end-to-end - I wanted full control over the infra, the data, and what the agents can touch. Locked it down with VPN and firewall, then connected it to the channels I actually use day to day.
- Private VPS - I own the compute, nothing runs on someone else's cloud
- VPN + firewall so it's not just sitting open on the internet
- WhatsApp and Telegram - talk to agents like you'd message a person
- Slack for automations and internal triggers
- Gmail - agents that can read, draft and send mail
- Whisper for voice input - speak instead of type
- Web search and live browsing - agents that actually go and look things up
- Scheduled agents - things that run on their own without me triggering them
- Monitoring dashboard - see what's running, what failed, what was triggered
- Multi-model - not locked to one provider, routes to OpenAI, Anthropic or local models depending on the task
Second Brain
2026
Everything I read, build, think about and work on ends up in one place. Obsidian is the vault, Claude Code is wired into it, and OpenClaw agents write to it automatically. It's not a notes app anymore - it's a system that captures stuff without me having to remember to.
- OpenClaw agents write notes into Obsidian from chats, emails and browsing - capture happens on its own
- Claude Code is hooked into the vault - it reads context from my notes when I'm coding
- Search and retrieval across the whole knowledge base, not just keyword matching
- Notes sync between OpenClaw and Obsidian so agents know what I know
- Automated summaries and connections between notes surface things I'd otherwise forget