Julian Alves and dlt: when expertise meets simplicity

Julian Alves has spent twelve years in the trenches of software and data infrastructure. Starting as a Java developer, he quickly transitioned to platform engineering. Setting up cloud environments, configuring networks, and mastering Kubernetes.

"Organizations really struggle to have reliable data infrastructure", Julian notes. "Not a lot of people have the expertise to build a very good data infrastructure. It takes a long time to get there."

When a Dutch fintech brought him on as a freelancer to build their AWS environment, cloud work evolved into data engineering as he helped provision their Redshift instance and set up Airflow. This sparked his journey into the data world.

Six years ago, he co-founded a company with two partners, one a fellow "hacker" and the other a data scientist. Their boutique consulting team of 7-8 people tackled complex projects: rebuilding data warehouses with dbt, redeploying Airflow, and rewriting data pipelines.

Their client roster ranged from startups to enterprices, where they rebuilt part of the data infrastructure in Kubernetes for the data science team. Another major project involved a three-year engagement completely rebuilding a client's data infrastructure and data models from scratch to align with product, finance, and marketing needs.

Four years later, Julian started Builders with a clear philosophy:

"The best technical solution means nothing if the team that is going to run and own that solution can't maintain it after you leave the job."

The dlt discovery: simplicity that actually works

While helping a logistics startup that needed to track field operations and capture GPS data points, Julian evaluated the usual suspects:

Fivetran? Too expensive. "They couldn't afford the risk of getting a €1000 bill just to replicate a database. That was ridiculous."

Airbyte? Problematic. "It's a hard tool to use as open source because there's a lot of bugs. It's quite complex to deploy."

Then he found dlt, but initially bookmarked it and moved on. It was too new, with few stars.

A month later, still frustrated with the alternatives, he gave it a real shot. The tests worked without the flaky behavior he'd come to expect from other tools.

"What stood out was that dlt is so simple and flexible," Julian explains. "You can run it anywhere, as a serverless function, a container, an Airflow DAG. You can deploy it using GitOps workflows and CI/CD."

Even better? "It behaves like a Python application completely, so it's also very easy to monitor using the tools that everyone knows for monitoring Python."

The open source advantage: real ownership

Julian's IT philosophy centers on open source whenever possible. Why?

"Open source tends to have a big community. It's easier to get help, easier to hire people, and these solutions tend to last for many years."

He points out that despite complaints, Airflow continues improving because so many companies are using and investing in it. "It will never go away unless something much better comes along."

What keeps dlt in his toolkit? Its elegant simplicity.

"The design of dlt is so elegant and simple that it might just work even with junior developers," Julian notes, something remarkable in the complex world of data infrastructure.

What makes Julian different as a consultant? He bridges technical possibilities and business needs:

"I see myself a bit like a bridge between the technical possibilities and the business needs. With every client, I first focus on understanding their actual data challenges and their goals before recommending any solution."

This pragmatic approach means sometimes using dlt, sometimes other tools, whatever best fits the client's needs and capabilities.

Ultimately, Julian became active on the dlt community, and contributed back. As dlt team was looking for extra help, and Julian had already worked with the dlt team on OSS contributions, we ended up hiring him for a number of projects.

The future: infrastructure that delivers value, not burden

Julian sees great potential in the shift toward open lakehouse architectures with Apache Iceberg. "It's something really interesting. A lot of companies will benefit from a more open source approach toward lake house and data warehouse solutions."

The market opportunity is enormous, and dlt is well-positioned to capitalize on it with features like native Iceberg support.

"Infrastructure should be boring. Python should be enough. With dlt, it finally is."

Building solutions that outlast your involvement

Julian is excited about working more closely with dlt. Why? It aligns perfectly with his core belief:

"Infrastructure should deliver value, not become a maintenance burden."

"I'm really excited about partnering more closely with dlt because I see a lot of potential in helping companies build infrastructure that delivers value instead of becoming a maintenance burden."

He appreciates dlt's approach: "The elegant design of dlt and this approach to be a Python-first solution is really incredible."

What stands out most about dlt is its "simplicity and documentation." Julian believes it will empower companies to own their data solutions much more easily than with tools like Fivetran or Airbyte.

If you want a pragmatic partner who builds solutions your team can actually maintain, Julian can help you take your dlt project to success.

Get in touch with Julian ("builders;"), here.

👉 Join the community