Inside NVIDIA’s AI Infrastructure for Self-driving Cars
summary
Project MagLev is NVIDIA’s end-to-end AV AI platform that ingests TBs/hour from fleet vehicles, stores multi-PB datasets, and orchestrates training across thousands of GPUs (DGX plus Constellation and Kubernetes/Apollo tooling). It builds data lakes, a curation/labeling stack, and simulation plus replay validation; active learning and targeted similarity search rank frames for labeling and give 3–5x improvement in useful sample selection. The platform runs in production and closes the loop from ingest to validated software builds.
AI-written summary · about
chapters (16)
- 0:00Intro
- 0:19The problem of selfdriving cars
- 1:43NVIDIAs principles
- 3:20Development workflow
- 5:59Platform overview
- 8:03Data lake overview
- 13:27Collection
- 14:04Active Learning
- 15:53Targeted Learning
- 17:11Ground Truth
- 18:29Labelling
- 19:35Examples
- 20:19Labeling
- 20:47Traceability
- 21:45Export
- 22:42Data Preparation
related talks
32mInside NVIDIA’s AI infrastructure for self-driving cars
28mMagLev: A Production-grade AI Platform Running on GPU-enabled Kubernetes Clusters
20mAccelerating AI pipelines w/ NVIDIA on AWS: An autonomous driving case
21mML Infrastructure for Autonomous Vehicles
05mInterview
40mFueling the AI Revolution with Gaming
42mAI and Self-Driving Cars
20mAI : Solving challenges and driving Innovation across Industries