I design and build complex, long-living systems where architecture actually matters.

Not CRUD apps. Not templates. Systems with feedback loops, state evolution, and real-world behavior.

• Architect high-load backend systems (.NET / PostgreSQL)
• Build modular monoliths that scale like microservices
• Design event-driven systems with deterministic cores
• Integrate AI safely (no “magic black box” logic)
• Develop simulation systems to validate product behavior

A system that models and optimizes human decisions through feedback loops

Core flow:

Signals → LifeState → Goals → Decisions → Actions → Feedback → Learning

Key ideas:

• Digital Twin (adaptive user model)
• Deterministic scoring engine
• AI as augmentation layer (not authority)
• Closed-loop learning system

Tech:

• .NET 8 / ASP.NET Core
• PostgreSQL + Event Sourcing
• CQRS + MediatR
• Local LLM (Ollama)

Synthetic environment that generates realistic signals for LifeOS

• Behavior modeling engine
• Scenario execution pipeline
• System validation under real-world conditions

Rebuilding a classic runtime with modern architecture

• Interpreter core
• Graphics pipeline
• Retro-computing experiments

• Terminal emulator (Win32 + Direct2D)
• SNES emulator (low-level CPU/PPU work)
• Redis-compatible in-memory cache
• Network tools & protocol experiments

• Deterministic core first
• Side effects at the edges
• Events over direct mutations
• State is a first-class citizen
• AI = constrained layer, never authority

• Systems > features
• Architecture > frameworks
• Clarity > cleverness
• Feedback loops > static logic
• Long-term thinking > quick hacks

• AI + deterministic systems fusion
• Personal intelligence systems (LifeOS)
• Simulation-driven validation
• High-performance backend architectures

I’m interested in:

• Complex backend systems
• AI-integrated products (with real architecture)
• Deep-tech / non-trivial engineering problems

• GitHub: https://github.com/dagrigorev