🇷🇺 Русский | 🇬🇧 English
Engineering Manager · Author of Nova Lang · Banking & Fintech · Remote across Russia
Started as a developer — 3D mobile games in C++, web services in Java, Go, C#, TypeScript, PHP, React, full stack. Moved into management and have been leading IT product development for 10+ years — from idea to production. Currently working in banking, heading distributed teams remotely, designing my own programming language Nova and using AI agents as part of everyday work.
| 10+ years | 5+ years | 5+ years |
|---|---|---|
| engineering management | web development · Java · Go · C# · TS · PHP · React | 3D mobile games · C++ |
«LLM writes the code — human reviews it». The first language designed for this era.
Algebraic effects make side effects visible right in the function signature. Reading a single line, a reviewer knows whether the function hits the DB, the network, reads the clock, or can fail — without reading the body. This changes the game for reviewing AI-generated code.
fn process_order(o Order) Db Net Time Fail -> Receipt
The signature alone says: the function uses the database, makes network requests, reads the clock, and can fail — and nothing else.
- ⚡ Effects + handlers — tests without mocks, the exact same code in production and tests
- 🔀 No
async/await— structured concurrency, no «function colour» - 🛡️ Contracts from Go to F* — a gradient from script to verified code in one language
- 🎯 Capability security —
forbid Net, Fsblocks an effect at compile time - ⏱️ Realtime blocks — the compiler guarantees no GC pauses for audio/trading/embedded
- 🔄 Time-travel debugging — a natural consequence of handlers
Looking for partner companies for contribution, sponsorship and joint research. Happy to discuss any format of collaboration — from one-off patches to a dedicated team.
Open to engineering leader roles — especially with teams where AI tools and modern engineering practices are already part of the culture.
- Scaffolding new services from scratch
- Refactoring legacy code without breaking business logic
- AI code review: SOLID violations, vulnerabilities, code smells
- Tools: Claude Code, Cursor, GitHub Copilot
- Generating unit and integration tests directly from code
- Building test cases and checklists from requirements and user stories
- Edge-case and boundary condition discovery
- Generating test data, mocks and stubs
- Designing microservice solutions with explicit trade-off analysis
- AI as a second voice in architecture sessions
- Writing Architecture Decision Records (ADR) with the agent
- Load analysis, bottleneck and failure point detection
- Breaking down business requirements into technical tasks
- API contracts, integration diagrams and data flows
- Generating documentation: README, runbooks, API descriptions
- Risk analysis and system degradation scenario planning
- Analysing existing systems and making improvement recommendations
- Comparing tech stacks with banking and CII compliance in mind
- Preparing technical presentations for stakeholders
- Forming vendor and off-the-shelf solution evaluation criteria
AI agents: Claude Code · Cursor · GitHub Copilot Languages: Java · Go · C# · TypeScript · PHP · C++ · Rust · SQL Frontend: React · TypeScript Databases: PostgreSQL · MSSQL · MySQL · Oracle · Kafka Frameworks: Spring Boot · Spring Cloud · Hibernate DevOps: Docker · Kubernetes · OpenShift · GitLab CI · Jenkins Processes: Scrum · Agile · Code Review · CI/CD · SDLC
| Metric | Before | After |
|---|---|---|
| Release cadence | 3 per year | 4 per quarter |
| Contact centre load | baseline | −20% |
| Time-to-Market | baseline | −20% |
| Critical bugs in prod | recurring | 0 |