Lead backend developer & system architect with nearly 7 years of commercial experience across adtech, fintech and retail — building high-load distributed systems and web applications on .NET. I take products end to end: architecture and domain design, monolith decomposition into microservices, integrations with external services, developer platforms and shared libraries. Beyond the classic backend, I design agentic and AI-assisted systems — the kind of project where LLMs become a production tool, not a demo — and keep a high bar for code quality: AI-generated and human code alike stays clean, consistent and review-ready.
Location:
Russia
- GitHub
- Telegram
- Email: [email protected]
- Russian — native
- English — B1 (intermediate)
- Languages:
C#,TypeScript,Rust,C++ - Core:
.NET,ASP.NET Core,EF Core,Microservices,System Architecture,DDD - AI & Agents:
LLM orchestration,agentic workflows,MCP,RAG,multi-agent systems,agent graphs (voluta),Claude Code,opencode,Orca - Data & Messaging:
PostgreSQL,Aerospike,Garnet,Redis,Kafka,ClickHouse,Elasticsearch,RabbitMQ,MongoDB - Infra & Ops:
Kubernetes,Docker,GitLab CI,Prometheus,Grafana,VictoriaMetrics,Keycloak
Full stack inventory
- Frameworks & UI:
.NET Framework,.NET Standard,ASP.NET MVC,Avalonia UI,WPF,React,Node.JS,Bun,Electron.js - Libraries:
EF Core/6,Linq2DB,Dapper,MassTransit,Benchmark.NET,xUnit,Moq/AutoMoq,OpenTelemetry,MiniProfiler - Code Quality:
custom Roslyn analyzers,Roslyn source generators,.editorconfig,analyzer rule sets,code style enforcement - Databases & Storage:
PostgreSQL,MongoDB,Aerospike,Garnet,Redis,ClickHouse,Minio,Amazon S3 - Messaging & Streaming:
Kafka(native client; consumer groups, partition replication, retention log),RabbitMQ,MassTransit - Monitoring & Profiling:
Prometheus,Grafana,VictoriaMetrics,Elasticsearch,Kibana,Logstash,Loki,Zabbix - Web/Application Servers:
nginx,traefik,IIS,self-hosted - Orchestrators & Runtime:
Kubernetes,Docker Swarm,Podman,Docker - Tooling:
Jira,Confluence,TeamCity,BitBucket,GitLab,GitHub,Draw.io,Postman,Keycloak - OS:
Windows,Linux (Ubuntu/Fedora/CentOS) - Methodologies:
SCRUM,Kanban,SOLID,Clean Code,ACID,TDD,DDD
Jan 2020 - Jun 2021, Energosfera
- Project Role - Middle FullStack Developer
- About Company - Largest retail company in the north in Russia
- About project/s - Internal CRM for management documents and transport logic
- Team size - 1 Project, 1 Tech Lead, 1 Team Lead, 2 - Product Managers, 8 - Developers
-
- Architected and developed three core internal systems: a bonus/loyalty system, cash register modules and a document routing service supporting key retail operations.
- Developed and maintained features for a legacy CRM and internal desktop applications (Electron.js), including integrations with retail hardware (cash registers, terminals).
- Customized and supported CI/CD pipelines, networking and reverse-proxy configuration for development teams.
- Implemented features required to comply with changing retail legislation, keeping company operations compliant.
-
C#,.NET Core,.NET Framework,Dapper,EFCore,JS/TS,React,Electron.JS,Docker,Prometheus/Grafana,Zabbix,PostgreSQL,Redis,RabbitMQ,Elasticsearch,Kibana
Jul 2021 - Feb 2022, Meridian Payments
- Project Role - Senior Backend Developer
- About Company - Fintech startup building a card-not-present payment gateway and billing platform
- About project/s - Payment processing, recurring billing, fraud detection and acquiring integrations
- Team size - 1 Tech Lead, 2 Product Managers, 6 Backend, 2 Frontend, 1 DevOps
-
- Designed and built the core payment gateway (.NET Core, PostgreSQL) processing card-not-present transactions with idempotency keys and exactly-once delivery semantics, shipped on Kubernetes with full OpenTelemetry tracing and Prometheus metrics.
- Implemented a real-time anti-fraud pipeline (Kafka, Redis) — heuristics + velocity rules that cut chargeback rate significantly within the first quarter.
- Built the recurring billing engine handling subscriptions, proration and dunning flows for SaaS merchants.
- Integrated with external acquirers and PSPs (3-D Secure, tokenization, webhooks) — standardized adapter layer so new providers shipped in days, not weeks.
- Owned PCI DSS compliance for the backend layer — no plain PAN storage, strict key management, audit-ready logging.
-
C#,.NET Core,EF Core,PostgreSQL,Kafka,Redis,Docker,Kubernetes,Prometheus/Grafana,OpenTelemetry,xUnit
Mar 2022 - For now, Hybrid.ai
- Project Role - Lead Senior Backend Developer, System Architect
- About Company - Marketing company, ranked 3rd in the world. Provides a complete package of necessary for opening and advertising on sites
- About project/s - CRM, DSP, SSP, Big Data statistic, more feature for marketing life, integration with popular solutions and advertising world supplier
- Team size - 1 Project, 4 Team Lead, 1 - System Architect, 8 - Product Managers, 16 - Backend, 14 - Frontend
-
- Conceived and built the product's MCP layer: an MCP gateway exposing the API as agent tools, generated from the API clients by a Roslyn source generator I wrote — replacing 88 hand-written wrappers; any drift between API and agent surface now fails the build.
- Designed the agentic contour end to end: agent runtime (CLI, web dashboard, audit trail), skill system, workflow orchestration and quality gates — plus an in-house orchestrator that turns Jira tasks into code with self-verification loops and upfront dollar-cost estimates. The harness layer runs on voluta — a LangGraph-style agent-graph engine for .NET I authored (open source); two production harnesses are built on it.
- Decomposed a 7-year-old core monolith into a modular core + BFF system with deep DDD and a plugin system; product teams extend it through a multi-agent toolchain (Claude Code, opencode, the in-house coding agent) orchestrated in Orca, while engineering owns the surrounding fleet of 20–30 microservices.
- Designed and built a notification service handling 50k rps at launch with horizontal scaling headroom, and custom DSL parsers letting teams express filtering, rules and domain queries declaratively; maintained the internal shared libraries every product builds on.
- Introduced AI-assisted development into daily delivery, backed by guardrails for agent-generated code — custom Roslyn analyzers, strict
.editorconfig, rule sets and a self-verifying verification harness (hermes). - Managed a team of 4 refactoring legacy modules and extracting satellite microservices; built an internal job orchestration & monitoring system (web UI, Keycloak auth), introduced an exception hierarchy splitting business vs technical errors, owned observability (Grafana/Kibana, VictoriaMetrics, Elastic) and set up Kubernetes production/dev environments with monitoring.
-
C#,.NET Core,.NET Framework,Dapper,EFCore,Linq2Db,MassTransit,Kafka,Redis,RabbitMQ,JS/TS,xUnit,MSUnit,Benchmark.NET,React,Docker,Kubernetes,Nginx,IIS,JMeter,Ansible,Loki,Proxmox,Amazon S3,Minio,Prometheus/Grafana,VictoriaMetrics,Zabbix,PostgreSQL,MongoDB,Elasticsearch,Kibana,Keycloak,Claude / LLM,MCP,agentic orchestration,Roslyn source generators,voluta,opencode,Orca,Bun,TypeScript
-
voluta — open source — my own agent-graph engine for .NET: LangGraph-style, but .NET-native. Stateful, cyclic, durable graphs with typed channels, Pregel-style supersteps, checkpoints, streaming and human-in-the-loop. Published on NuGet under my org dot-stbl.
- Cycles not DAGs; a thread can pause for days and resume in another process.
- Two production agent harnesses at Hybrid.ai are built on it.
-
notifliwy — open source — my own .NET library: a distributed handler & pipeline for notifications, built for event-driven architectures with a fluent builder API. Published under my org dot-stbl.
- Composable processing pipeline (input → connector → conditions → mapper → exporter), fully async with DI integration.
- Pluggable transport providers (Kafka / MassTransit), built-in metrics and OpenTelemetry diagnostics.
-
Real-time crypto analytics & trading platform — private, self-hosted — a personal platform that ingests and processes ~150k events/sec from 17 exchanges and exposes unified market analysis and automated trading.
- High-throughput event pipeline with low-latency aggregation across heterogeneous exchange feeds.
- Source kept in a self-hosted Git; not public — details and walkthrough available on request.
-
Self-hosted home ecosystem — private — a suite of personal services and automation I design, build and run on my own infrastructure.
Nefteugansk's Technology College
- Faculty - Production
