LLM agents in real workflows, on top of Go backends that hold real load.

Architected and built the Transport Management System (TMS) from the ground up: 13 microservices and 2 API gateways.

• Designed the system architecture from scratch — service boundaries, contracts, data flow, event topology — as founding backend engineer; mentored and onboarded the team from 1 to 4 engineers, including Java → Go transitions.
• Built 13 microservices using DDD, including Vehicles, Drivers, Logistic Points, Carriers, Trips, Mobile Backend, Notifications and Locator (real-time driver coordinates via HTTP streaming), plus admin and mobile API gateways.
• Reduced Trips response time from 1,800 ms to 150 ms (12× improvement) through query denormalization, index tuning, and table flattening — significantly improving dispatcher operations.
• Introduced gRPC for service-to-service communication with disciplined schema management, reducing inter-service latency and improving contract reliability.
• Kafka-based event flow with idempotency and inbox/outbox patterns to guarantee exactly-once semantics across distributed workflows.
• End-to-end observability with OpenTelemetry, Prometheus, Grafana — dashboards tracking business KPIs (trips dispatched, driver utilization) and application metrics (p99, error budgets).
• Optimized GitLab CI/CD and Kubernetes deployments with Helm — ~40% faster deployment time, standardized release practices.
• Defended architectural proposals at the company-wide Architecture Review Board — all approved and rolled to production.
• Interviewed 50+ Go candidates; mentored mid-level engineers through technical growth and code review.

AI engineering

• Built a shared Claude Code workspace used by all 8 developers on the team, across ~15 Go microservices, the frontend, the mobile app and e2e tests: root and per-service AGENTS.md, team rules for contracts, migrations, Kafka, caching, tests and the transport layer.
• Connected agents to Jira, GitLab, Figma, Grafana and Sentry through MCP.
• Automated the task flow end to end: Jira ticket to agreed spec and plan, implementation with linters and tests, self-review, MR with a review summary attached.
• Agent code review against team rules: a scheduled watcher picks up review requests from the team channel; a second workflow checks reviewer comments against the code, applies agreed fixes and replies in the threads.
• Agents draft Jira tasks from the code, the admin UI and the designs: description, acceptance criteria, subtasks.
• Operations access with guardrails: log search in Grafana / VictoriaLogs, read-only Kubernetes with explicit limits for dev, uat and prod. The workspace has its own CI that tests skill scripts and lints instructions.

Go · gRPC · Kafka · PostgreSQL · Redis · Kubernetes · Helm · GitLab CI/CD · OpenTelemetry · Prometheus · Grafana

Designed and delivered high-load backend services for one of the largest grocery delivery platforms in Eastern Europe.

• Designed and built the SMS notification service at ~50,000 RPS — customer transactional messaging across millions of users; fault-tolerant with retry, deduplication, and provider failover.
• Built the user reviews service handling ~2,600 RPS with thousands of reviews written daily; designed for read-heavy traffic with Redis caching and PostgreSQL partitioning.
• Built the couriers microservice supporting real-time courier operations across the delivery network.
• Modernized the legacy monolith by extracting bounded contexts into independent microservices, raising quality through unit, integration, and functional test coverage.
• Defended service designs at the architecture committee and led production rollouts.
• Participated in production on-call rotations and led post-mortems for high-severity incidents.

Go · gRPC · Kafka · PostgreSQL · Redis · Docker · Kubernetes · GitLab CI/CD

Built and operated backend services for two business-critical domains: customer reviews and warehouse management.

• Engineered the customer reviews service at ~8,500 RPS across millions of products, with Elasticsearch for storage and full-text search.
• Designed and implemented anti-fraud mechanisms against review manipulation and rating inflation — combining behavioral signals, rate limiting, and pattern detection to protect platform trust.
• Built the Warehouse service for operational fulfillment workflows, using Go, NATS, MongoDB, Redis, and gRPC.
• Worked with NoSQL and relational databases at production scale, with focus on read/write performance tuning.

Go · gRPC · NATS · MongoDB · Elasticsearch · PostgreSQL · Redis

Backend developer on enterprise outsourced projects for high-load clients.

• Tutu.ru (largest Russian travel & ticketing platform) — part of a 4-engineer backend team responsible for bus carrier integrations. Built and maintained 10+ integrations with European and CIS bus operators (Lux Express and others) over a mix of REST and SOAP, normalizing diverse carrier interfaces into a unified internal booking model.
• Transitioned to Go on internal Simbirsoft projects — first hands-on production Go experience, marking the start of full specialization in the Go ecosystem.

PHP · Laravel/Lumen · Go · REST · SOAP · MySQL · PostgreSQL

Before focusing on hands-on engineering, held progressively senior IT leadership and management roles in regulated industries:

• Head of Software Development — Magazin Vnimaniya (regional media holding, including Kommersant-Kuban) · 2016–2018
• Head of IT Department — Kuban Universal Bank · 2010–2014
• Head of Information Security & Operations — Kuban Universal Bank · 2008–2010
• Director — IT Standard LLC · 2006–2008
• Earlier roles in banking IT and operations from 2000.