Using AI and Computer Technologies to Enhance IT Infrastructure
summary
Argues that industrial AI works by combining on‑premise industrial training servers with cloud or data‑center scale GPUs (nvidia a100) plus low‑latency storage and melanox networking so GPUs with 6912 parallel lanes stay fed; recommends starting with an MVP on small curated data, then develop, train, infer at scale. Describes a working POC environment and industrial‑rated servers for harsh sites, and deploying trained models to edge units from jetson to xavier for robotics, drones, predictive maintenance and process control.
AI-written summary · about
chapters (9)
- 0:00Start
- 1:00Data is like Energy
- 1:50The Data & Service Lifecycle
- 5:21Growth of Neural Networks, Deep Learning and Classic ML
- 6:29Building a Use Case
- 8:33Computational Power
- 12:27Deployment Models
- 13:32Manufacturing Use Cases
- 15:12Summary
related talks
05mAI and ML are changing traditional business
40mAI Factory - Enabling Scalable Innovation & Intelligent Systems
40mFueling the AI Revolution with Gaming
06mAI Inference at the Edge: Beyond the Hype
26mAI will drive the industrial IoT to the cloud
31mScaling AI for Industry
29mFrom Experiment to Enterprise: Architecting AI for Stability and Scale
27mAI at Scale in the Enterprise