Open to opportunities

Denis Ilin

@denisilin

Message

Senior AI Infrastructure Engineer building scalable LLM inference and agentic systems for high-load production AI.

Zimbabwe

Message

What I'm looking for

I’m looking to build production-grade LLM/VLM inference and agentic systems—scalable, observable, and cost-efficient. I want to own end-to-end architecture, automate GPU infrastructure, and deliver workflow automation that integrates cleanly with business systems.

I’m a Senior AI Infrastructure Engineer with 9+ years of experience specializing in scalable LLM/VLM inference and multi-agent systems. I focus on delivering reliable, high-load production AI services with cost-efficient infrastructure and measurable reliability.

I’ve architected agentic workflows with LangGraph to automate complex business processes, and built production ML inference services with deep observability. My work includes self-hosted LLM/VLM inference using vLLM, plus performance optimizations like continuous batching, KV-cache optimization, and VRAM management.

I also drive automation across the stack—implementing RAG with centralized knowledge-base patterns, integrating LLM APIs into corporate infrastructure, and building automated GPU provisioning and lifecycle management. Across backend and AI platforms, I prioritize dependable execution, workflow automation, and continuous improvement through system monitoring and feedback loops.

Experience

Work history, roles, and key accomplishments

Current

LLM Infrastructure Engineer (R&D)

Current

PropellerAds

Oct 2023 - Present (2 years 9 months)

Architected and implemented agentic workflows with LangGraph and deployed production LLM/VLM inference services integrated into corporate infrastructure. Built observability and RAG systems and operated a self-hosted vLLM inference stack with automated GPU provisioning.

LangGraph Multi Agent Systems LLM Inference VLLM RAG PostgreSQL Docker Observability Queue Depth Kafka Embeddings Workflow Automation Python fastAPI

Senior Backend Engineer

Elama

Jun 2021 - Oct 2023 (2 years 4 months)

Re-architected core billing and commission engines using Domain-Driven Design to ensure consistent financial operations. Migrated legacy monolithic modules to microservices to improve system scalability and deployment reliability.

Domain Driven Design Microservices Backend Engineering Billing Systems