- Lead the architecture program extracting the wallet, slots and bonus subsystems out of the monolith: target balance model, bet/win routing, bonus-split and regulatory rules, phased migration plan — researched, designed and defended to the CTO; now driving phase-1 implementation of the dedicated slots service.
- Own the payment domain: deposits and payouts across 6+ crypto providers and PSPs, incl. TON/ton-connect and p2p terminal flows. Designed the terminal-payments subsystem end-to-end (schema-driven provider contracts, admin tooling, per-terminal allowlisting) — new integrations ship with no code release.
- Established and drive the mandatory review gate for every money-touching MR across both Python backend teams (~11 engineers), plus a production-SQL validation gate.
- Security & incident response: ran an internal security pass alongside external pentesters — identified, proved with PoCs and remediated critical vulnerabilities in payment and authentication paths ahead of a brand launch; led root-cause forensics of an active intrusion under live money flow (cross-system log analysis, access chain traced and closed).
- Resolved production incidents to root cause under live money flow: a memory leak (1 GB→4 GB), deposit double-credit races, a transaction-integrity bug silently dropping DB writes; cut a 14s full-scan payment query with a targeted partial index.
- Rolled out OpenTelemetry tracing and self-hosted Sentry into the shared core library used by all company microservices; delivered the backend for a white-label brand launch (branding, currency, payments, go-live).
- Built and rolled out a multi-agent AI code-review pipeline (Claude Code, MCP, hooks), adopted by the team via an internal plugin marketplace.
Nikita Kurinnyi
Summary
Backend engineer, 10+ years of production Python, building high-load transactional systems across FinTech, iGaming, HealthTech and GameDev: payment routing, idempotent money callbacks, multi-provider (PSP + crypto) integrations.
Architecture owner beyond senior scope: currently leading the monolith→microservices extraction program of a real-money platform (wallet, slots, bonus domains) — target balance model, phased migration plan, decisions defended directly to the CTO.
Led backend teams of up to 5; set code-review, hiring and CI/CD standards from scratch; drive money-critical review and application-security practice across two backend teams (~11 engineers).
Bring agentic AI into production workflows, not demos: multi-agent code-review pipelines and MCP-based tooling (Claude Code plugins/hooks) adopted by the team; build full backend systems end-to-end through an agentic LLM workflow.
Skills
- Languages
- Python, Go, SQL, Bash
- Backend
- FastAPI, Django, asyncio, Celery, SQLAlchemy, Alembic, pytest; gRPC/Protobuf, GraphQL, REST
- Payments
- multi-PSP & crypto-provider routing, idempotent callbacks, reconciliation, TON/ton-connect, p2p terminal flows
- Data & Messaging
- PostgreSQL, ClickHouse, Redis, BigQuery; Kafka/Redpanda, RabbitMQ; PySpark, ETL
- Infra & Observability
- AWS, Docker, Kubernetes, Terraform, GitLab/GitHub CI/CD; OpenTelemetry, Sentry, Grafana, Prometheus
- AI Engineering
- agentic pipelines in production (Claude Code MCP servers, hooks, plugins), multi-agent code review; RAG on pgvector, embeddings, OpenAI/Anthropic APIs; LangGraph
Experience
- Architected and delivered an internal operations dashboard automating missing-payment investigations, cutting per-case time by 90%+ (~6 min to ~30–40 sec).
- Architected centralized credential management for RPA bots and polling services, improving reliability and failure monitoring.
- Designed a multi-layer analytics data model (raw, stg, dds, emart) for consistent BI and data quality.
- Integrated OpenTelemetry distributed tracing to speed incident response and surface bottlenecks (e.g. GraphQL N+1); championed a GraphQL→REST/FastAPI migration.
- Co-designed and led the core backend: 4 foundational microservices (Auth, User, Wallet, Transactions) on Python + gRPC/Protobuf.
- Engineered fault-tolerant payment integrations (Stripe, National Bank of Kuwait) with robust idempotency for transaction integrity.
- Established engineering foundations (CI/CD, code-review standards, stack choices) and took part in backend hiring.
- Introduced a load-testing practice (Yandex.Tank) and capacity baselines in CI; built end-to-end observability (OpenTelemetry, Jaeger, Grafana, Sentry).
- First backend hire; built AWS cloud infrastructure from scratch for an AI chat-companion app.
- Cut the AI response pipeline (~10 model calls) p95 latency by 50%+ (~5s to ~2.2s) with the ML team.
- Architected the data layer: DynamoDB for low-latency chat history and embeddings, BigQuery for analytics.
- Designed and shipped a real-time analytics pipeline (FastAPI, Kafka, PySpark, BigQuery) processing 250M+ events/day.
- Delivered critical hourly aggregates for reporting and product analytics; led sprint planning, roadmap and code review.
- Led backend for Tochka XS, a near zero-touch SMB banking product.
- Introduced test coverage, Prometheus monitoring and Sentry into a legacy system; built core services on Python/asyncio + PostgreSQL with complex internal-banking integrations.
- Built a high-performance CS:GO replay parser feeding ML/analytics; maintained real-time Dota 2 feeds for live broadcasts.
- Built third-party API integrations and a user notification system; helped decompose a Django monolith using Celery + Redis.
- Shipped backend services for startups and clients (Django, databases, caching, queues).
Projects
Personal systems built end-to-end through an agentic LLM workflow (Claude Code), strict hexagonal + DDD + CQRS architecture: a social-poker Telegram Mini App platform and an agent knowledge/memory system. Stack: Python, FastAPI, SQLAlchemy (async), PostgreSQL/pgvector, Redpanda (Kafka), Temporal, React.