

Scaling ML Infrastructure Like a Pro with Elliott Clark
Apr 16, 2025
Elliott Clark, a DevOps expert and founder of Batteries Included, shares his vast experience in scaling machine learning infrastructure. He discusses the critical role of performance testing, emphasizing the use of production traffic for safe and effective code testing. Elliott also highlights the importance of a supportive team culture in navigating microservices and scaling challenges. He advocates for open-source solutions to aid mid-sized companies, and introduces a unique pricing model that democratizes access to robust infrastructure, making it easier for smaller setups.
AI Snips
Chapters
Transcript
Episode notes
Choosing Infrastructure
- Choose default infrastructure recommended by the community (e.g., Postgres for databases, Let's Encrypt for SSL).
- Identify pain points before exploring niche solutions, scaling based on actual needs.
Scaling Challenges
- Scaling is not just about adding virtual environments; it involves scaling people and their understanding.
- Scaling requires teaching teams to navigate codebases and handle increasing complexity.
Scaling Teams
- Foster a culture that encourages learning, experimentation, and knowledge sharing.
- Reward teaching and create a safe space for engineers to try new things without fear of failure.