Looking for a job

Davlat Magzumov

@davlatmagzumov

I build high-performance backend, ML, and data systems with C++, Python, GCP, and Kubernetes.

I've built high-performance backend, ML, and data systems at Yandex, Huawei through Coleman, Synergetics.ai, and escs.io.

At Yandex, I optimized C++ advertising model pipelines to reduce training time by 25% and increase serving throughput by 30%. I also scaled Yandex Maps data pipelines, improving throughput by 25% for six downstream services.

I refactored time-critical ROS2 components to cut average task latency by 80ms, and built C++ and Python packages that improved autonomous-system task success by 15%.

More recently, I've delivered backend microservices, reduced P95 latency by 30% with Redis caching, migrated critical analytics datasets to BigQuery, and built an AI-driven Playwright test suite integrated into CI/CD.

Experience

Work history, roles, and key accomplishments

Current

Software Engineer (Middle)

Current

Escs.io

Apr 2026 - Present (6 months)

Developed an AI-driven E2E test suite using Playwright and a fine-tuned LLM, integrated into CI/CD. Automated data pipelines on GCP, migrating datasets to BigQuery and authoring complex SQL for BI reporting.

Python NodeJS BASH Kubernetes GCP Playwright LLM CI CD BigQuery SQL Data Pipelines

Software Engineer (Middle)

Synergetics.ai

Dec 2025 - Mar 2026 (3 months)

Cut P95 backend latency by 30% via multi-layer caching and query optimization. Designed and deployed 5 core microservices using Docker and API-first approach.

Python NodeJS BASH Kubernetes GCP Redis Microservices Docker API Design

Backend Developer / System Administrator

Coleman (Huawei)

Mar 2025 - Aug 2025 (5 months)

Architected core components for GaussDB query processing layer, enabling serverless transition. Developed C++/Python automation tools reducing deployment time from over an hour to under 30 minutes.

C++Python Linux Docker Kubernetes BASH GaussDB Automation

Software Engineer (Backend & ML)

Yandex

Jun 2022 - Feb 2025 (2 years 8 months)

Optimized C++ advertising model pipelines, reducing training time by 25% and increasing serving throughput by 30%. Redesigned feature retrieval, accelerating data access 3x. Built data pipelines for Yandex Maps and refactored ROS2 nodes to cut latency by 80ms.