I'm Roman, Technical Founder & AI Product Engineer.
After 10+ years in product & growth, I went all-in on building AI-powered products. I understand business logic and build like an engineer.
Flutter Dart TypeScript Python PostgreSQL Deno Astro Claude OpenAI Gemini DeepSeek Qwen Cursor Lovable
Supabase Appwrite Docker Cloudflare GitHub Grafana Zabbix Metabase VictoriaMetrics Sentry PostHog Google Analytics Yandex Metrika Langfuse LiteLLM LightRAG Neo4j Qdrant Flowise n8n Hetzner OVHCloud Contabo Netcup Vaultwarden fail2ban Let's Encrypt GuardrailsAI LLM Guard Outline Linear Notion Miro Obsidian Joplin Figma Telegram Slack Zoom Google Meet NTFY Backblaze Yandex Cloud Playwright Crawl4AI Brave Search Bubble.io FlutterFlow Tilda
AI-first travel planner — itinerary generation, POI scoring, real-time recommendations, and an AI chat assistant that knows everything about your trip.
Tech blog about AI engineering — deep dives into LLM orchestration, edge functions, context engineering, and building production AI systems. Written from the trenches.
Use Gemini CLI from Claude Code to review code and delegate tasks.
Interactive visual map of your Claude Code ecosystem: skills, agents, MCP servers — auto-discovered from .claude/ config.
Multi-agent AI consultation framework for Claude Code via MCP — second opinions from OpenAI, Gemini, Qwen, DeepSeek.
Self-hosted runtime for durable AI agents — fractal multi-agent orchestration, 6-layer memory, MCP tools, local dashboard.
- Context Engineering: How to Manage Context for AI Models and Agents
Practical framework for structuring context — 6 layers, 4 strategies, persistence, and subagents with real Claude Code examples. - Multi-Provider LLM: How to Stop Depending on a Single API
Architect resilient LLM systems with fallback chains, task routing, and cost balancing across providers. - Circuit Breaker in Deno Edge Functions
Build resilience layers for edge functions with circuit breaker patterns, retry logic, and fail-open rate limiting. - LLM Observability with Langfuse
Enable tracing, prompt management, cost tracking, and evaluations with self-hosted Langfuse. - Maestro + Flutter: E2E Tests in YAML Without the Pain
Production patterns for mobile testing — regex matchers, iOS quirks, and AI-powered assertions.