Can SREs become AI/ML engineers?
summary
Kendra runs SRE teams that also build and operate AI platforms for 55,000 delivery engineers, arguing SREs can become ML engineers by owning MLops: data pipelines, CI/CD, observability and governance across vector DBs, JupyterHub, KServe/NVIDIA Triton and private or sovereign clouds. A NanoGPT experiment trained on four SRE books produces gibberish, proving data quality and indexing matter, so SREs add value via vector DB indexing and RAG pipelines and by building automation, agent orchestration and responsible-AI guardrails.
AI-written summary · about
chapters (12)
- 0:00Introduction to AI and SRE
- 0:51The Role of AI in SRE
- 1:20Introduction of the Speakers
- 1:40The Importance of Passion in SRE
- 3:43AI in Customer Services
- 4:57The Need for MLOps
- 5:51The Complexity of Modern AI Systems
- 7:19The Perfect AI Stack
- 12:02Challenges in AI and SRE
- 13:39The Dual Role of SREs
- 15:32Running AI Platforms
- 21:24Q&A Session
related talks
44mSRE and ML: Why It Matters
41mAIRe - AI Reliability Engineering
44mSRE for ML: The First 10 Years and the Next 10
24mAutonomous Application Management for DevOps and SRE
44mAgentic Observability — The Path to AI Co-SRE
36mTransformers in SRE Land: Evolving to Manage AI Infrastructure
46mIntroduction to AI engineering
46mI know Machine Learning but... am I ready for the real world?