Start Here

Why this guide

Data engineering and the neocloud industry intersect from two directions:

• From the outside in. A data engineer at an AI company runs pipelines that produce training data, feed inference, and govern cost. The decisions about which neocloud to use, how to move data to it, and how to instrument cost cascade into the DE's job.
• From the inside out. A data engineer at a neocloud builds the telemetry, billing, customer-usage analytics, and marketplace-event pipelines that run the platform. Without those data systems, the neocloud can't price, can't bill, can't detect fraud, and can't plan capacity.

This guide covers both. The two perspectives reinforce each other — once you've seen what a neocloud's internal DE team builds (Section C), you understand better why your data-engineering decisions as a customer (Section B) land the way they do.

The two perspectives

Section B is the practitioner playbook. Five chapters covering provider choice, training-data pipelines, orchestration, and cost engineering. If you're a DE at an AI company running workloads on Vast or CoreWeave or Together, this is the working manual.

Section C is the inside view. Five chapters on telemetry, billing, customer-usage analytics, marketplace event streams, and trust / fraud data. If you're a DE at a neocloud — or are interviewing at one, or want to evaluate one as a vendor — this is the working model.

Section D synthesizes. One chapter on what each side takes from the other.

Who it's for

• Data engineers at AI companies running production workloads on GPU clouds.
• Data engineers at neocloud companies (CoreWeave, Crusoe, Lambda, Together, RunPod, Vast, etc.) or considering joining one.
• Engineering leaders evaluating neoclouds as vendors and wanting to understand the data systems they're buying into.
• Anyone interviewing for a DE role at the boundary between AI compute and data infrastructure.

Reading order

Linear if you're new to the topic. Section B (practitioner) → Section C (inside) → Section D (synthesis). Each section can also be read standalone if you're already familiar with the other.

If you only have time for the most-load-bearing chapters, read 02 (Choosing a Provider) and 07 (Billing & Metering) — both are flagged as critical.

Prerequisites

Assumes general DE familiarity — SQL, batch / streaming pipelines, orchestrators, warehouses. Doesn't assume deep GPU or AI infrastructure background, but the surrounding context helps. The Neocloud topic (separate guides) provides company-by-company background; this guide assumes you can look up specifics there as needed.