Back

Alexey Golev

Ghost
alexeygolev.blog· irregular· Published 2 months ago

Production engineer Alexey Golev writes about the practical realities of software meeting real-world conditions, such as reliability of multi-agent LLM pipelines, ML model serving architecture, and token cost of parallel agents. Readers get experience-based lessons on applying distributed systems thinking to AI systems.

multi-agent systemsllm agentsagent reliabilitydistributed systemsmodel servingkubernetesml platformtoken costai agent cost optimizationproduction engineering
Platform
Ghost
Frequency
Irregular
Latest issue
Sep 4, 2026
Language
English
Written for
Backend, platform and ML engineers building LLM or ML systems in production
Tone
practical
Tracked since
Oct 25, 2026

Website: https://alexeygolev.blog/