Skip to primary content
Dr. Marcus Vance

Dr. Marcus Vance

Principal MLOps Engineer

Alumni of University of Cambridge

Dr. Marcus Vance leads MLOps infrastructure and distributed model serving at Esaholic. He specializes in GPU cluster optimization, vLLM throughput scaling, vector database indexing, and continuous evaluation pipelines.

Marcus holds a PhD in Computer Science from the University of Cambridge, focusing on distributed systems and parallel neural network execution. Prior to joining Esaholic, he engineered high-throughput inference backends handling over 50M daily requests.

Areas of Technical Expertise

MLOps & LLMOps vLLM & TensorRT-LLM Distributed GPU Clusters Vector DB Optimization Qdrant & Milvus Scaling Prometheus Model Monitoring

Credentials & Certifications

  • PhD Computer Science — University of Cambridge
  • NVIDIA Certified Deep Learning Professional

Talks & Keynote Presentations

Scaling Small Language Model Inference on Consumer Hardware

MLSys Europe • 2025