Looking for a job

Arsen Ozhetov

@arsenozhetov1

Backend engineer at YurtTech who standardized observability across 10 microservices and cut a ClickHouse query from 1,000 ms to 34 ms.

Kazakhstan

Message

At YurtTech, I work on a multi-tenant CCTV platform for real-time streaming, video analytics, and face recognition. I built a reusable Spring Boot OpenTelemetry starter that standardized metrics and distributed tracing across 10 microservices.

During a four-month acting technical lead assignment, I led architecture decisions for three backend engineers. I also fixed a ClickHouse query that lacked tenant filtering and optimized a measured query from 1,000 ms to 34 ms.

At Huawei Technologies, I increased data-pipeline throughput 2× by replacing polling and scheduled synchronization with real-time processing, and implemented a C++ UDP transport with Reed-Solomon erasure coding. I’ve also developed AgentShield, a Go circuit breaker for LLM agents, and contributed open-source fixes and translations.

Experience

Work history, roles, and key accomplishments

Current

Backend Engineer

Current

YurtTech

Jan 2026 - Present (8 months)

Led architecture decisions for 3 backend engineers during a 4-month acting technical lead assignment; standardized metrics and distributed tracing across 10 microservices by building a reusable Spring Boot OpenTelemetry starter. Restored retention cleanup, closed a cross-tenant data leak, and eliminated recurring ClickHouse out-of-memory crashes.

Java Spring Boot OpenTelemetry Clickhouse Kubernetes Docker Kafka RabbitMQ PostgreSQL Redis Qdrant Gitlab CI ArgoCD Helm

Software Engineer

Huawei Technologies

Jun 2024 - Aug 2025 (1 year 2 months)

Increased internal data-pipeline throughput 2x by consolidating background tasks and replacing cross-service polling with real-time processing. Restored file transfers on lossy networks by implementing a C++ UDP transport with Reed-Solomon erasure coding.

C++UDP Reed Solomon Codes Data Pipelines Real time processing

Backend Engineer

Arbuz Group

Mar 2023 - Nov 2023 (8 months)

Accelerated shelf-level remaining-stock calculation 7.5x and eliminated peak-load catalog timeouts by replacing brute-force queries with hash-map and prefix-sum algorithms. Reduced warehouse order-picking time by 20% by building a routing service with fixed serpentine paths, and reduced CI/CD duration 90% by selecting tests for changed code.

Python fastAPI Algorithms CI CD Test Selection Routing