Katkat
Back

Design teleoperations data models and pipelines

ZooxData Engineer · Onsite
MediumData EngineerOnsiteSystem Design

Last updated

This question evaluates competency in data modeling, real-time and batch pipeline architecture, telemetry ingestion, metrics design, and operational monitoring for teleoperations and autonomous vehicle stability.

You work on a fully autonomous robotaxi service. Passengers can request rides, get picked up, and get dropped off. During trips, the robotaxis are continuously monitored by tacticians in a teleoperations center.

Teleoperations tacticians can:

  • View a robotaxi's live location.
  • View real-time video of the vehicle surroundings.
  • Take control of or assist the robotaxi when a stuck event or other help request occurs.

The operational flow is:

  1. A robotaxi requests help from teleoperations.
  2. The request is placed into a queue.
  3. A tactician is assigned to service the request.
  4. The tactician may observe, advise, remotely control, escalate, or resolve the issue.

There are always more active robotaxis than tacticians. One tactician may support multiple robotaxis over a shift.

Design the following:

  1. A data model to monitor the efficiency and effectiveness of the teleoperations function.
  2. A data architecture and pipeline to measure vehicle stability across the autonomous fleet.

For the teleoperations model, include the core entities, table grains, relationships, and metrics. For the vehicle stability pipeline, describe data sources, ingestion, storage, transformations, quality checks, metrics, and downstream consumers.