Blog

A data platform should run like a power grid: routine faults cleared automatically inside a bounded scope, anything risky escalated to a person. This is the four-level autonomy ladder dltHub is climbing, and the context graph that makes each step safe to take.

In March 2024, we published “dlt adds Reverse ETL”. This piece walks through an example GTM workflow: turning a hiring signal into enriched, sales-ready company and people records, then reverse-ETL them to Attio.

In previous blogs we talked about how to build canonical models. This blog takes that further: we declare a BigQuery property graph over those models, query it with GQL, then ask it questions in plain English. The ontology resolved in the first step is what makes the answers right at the end.

An agent can write a pipeline now, faster than a contractor and cheaper than a hire. So the question is obvious: What people do we still need on the team?

In this article, I walk through moving a HubSpot connector and its dbt models from Fivetran to dlt. I keep the old stack running as the reference, build against its schema, and diff until every mismatch has an explanation.

dlt made agents good at building pipelines. dltHub makes their work safe to run: agentic alerts, team workspaces, and managed infrastructure for your whole team.

In regulated life sciences, "can you explain your data?" has moved from a QA question to a board-level one. dltHub turns ingestion into validated, testable code - built on dlt, run on dltHub - so your team and their agents produce data that lands in Snowflake with the lineage, tests, and evidence regulators expect.

dlthub ships 2 ways to run transformations - you can bring your own dbt project or use dlthub native transformations. This posts aims to explain to practitioners and technical decision makers how the two methods are different. Let’s dive in!

In this article, I'm exploring an AI-native architecture that aims to solve the traditional central data team problem.

dlt handles ingestion. dltHub adds the five ecosystem layers you'd otherwise build yourself: AI harness, context catalog, transformation, orchestration, managed infra.

Airbyte promises a self serve UI, a catalog of connectors, no code. Sounds great on paper, but in practice this vision is plagued by skill barriers, reliability issues, and high cost of ownership.