kolesnikovm

Designed and developed a message communication system between high-loaded services (∼500M messages/day). Reduced required hardware resources by ∼40% (2k 1.2k cores) through asynchronous processing. Reduced the number of production issues by 80% (10 to 2 per release) by implementing root cause analysis practice. Increased both code updates release and CI testing pipelines stability by ∼50% (2x fewer test failures). Created and led a team of 7 developers (intern to senior level) with zero turnover for over a year. Developed and negotiated new KPIs for Dev and QA teams, increasing feature delivery rate by 20% (15 to 18 features per sprint).

Designed highly available microservices, developed distributed algorithms, and supported the production environment. Migrated infrastructure (in-house servers to cloud), increasing system throughput by ∼4x (250k to 1M notifications/day). Improved calculation of courier travel route, allowing to fulfill ∼15% more orders by the same number of couriers. Optimized bulk notifications upload time by ∼50% (10k rows from ∼1m to ∼30s). Improved low latency search algorithm performance by ∼25% (2s to 1.5s), saving 3h per 20k client’s search requests. Developed real-time banking transaction processing system. Reduced tail latency for transaction processing by ∼15% (900µs to 750µs for 99th percentile). Increased auto-testing pipelines stability, reducing ∼90% failures and code delivery time by 80% (3h to 30m) for ∼40 developers. Designed and developed a low-overhead tracing system (∼3% CPU tracing 10% of requests). Optimized encoding algorithm, increasing system throughput by ∼25% (6k to 7.6k TPS). Developed highly available services and search algorithms for 5M+ DAU e-commerce platform. Optimized large baskets (300+ items) loading speed by ∼80% (10s to 1.8s), decreasing payment processing time by 40%. Designed and developed a stable RPA solution to regularly update 100+ suppliers’ data, ceasing manual uploads. Reduced order loading time latency by 60% (10s to 3s), saving 2h per 1k client’s requests. Reduced main dashboard loading time by 50% (3s to 1.5s), improving bounce rate by 30%.