From Databricks POCs to Building a Data + AI Community

[Brahmareddy]

Over the last few years, I have spent a lot of time experimenting with one question: How do we turn Data + AI capabilities into applications people actually want to use?

That curiosity led me to build multiple POCs and applications using Databricks + Lovable.

Some focused on recommendations and personalization. Some explored decision intelligence, enterprise agents, context engineering, HR experiences, governed conversational analytics, and other Data + AI patterns.

One of those projects, Future Movie Discovery, also received an Honorable Mention in a Databricks Hackathon.

But the biggest learning was not about any single technology.

It was seeing what happens when we connect: Data + Context + AI + a great application experience.

Databricks gave me the foundation for data, governance, intelligence, Genie, agents, and AI capabilities.

Lovable.dev helped me move those ideas into usable applications much faster. Interesting part is databricks also invested in Lovable for Data + AI apps capabilities.

After building many of these experiments and sharing what I learned, I was recently selected as a Lovable Industry Lead for Data + AI.

For me, this feels less like a new direction and more like the next step in the same journey.

Now I want to take it further.

I am working toward a broader Data + AI community initiative where we can continuously explore:

Data + AI updates, practical tutorials, real application architectures, lessons from POCs, new tools and platforms, builder stories, talks, meetups, community projects.

The goal is not just to talk about AI.

It is to help people answer:

What is possible with Data + AI?

What should I build?

How can I turn an idea into a useful application?

And most importantly:

What can we learn from people who are actually building?

I am currently exploring Data + AI Frontier / DataPlusAI.org as the direction for this broader community.

Databricks has been a major part of how I reached this point, so I wanted to share the idea here first.

I would love to hear from the community:

What would make a Data + AI platform like this genuinely useful for you?

Tutorials, architecture breakdowns, real POCs, product updates, meetups, talks, builder stories, or something else?

I am excited to keep building, learning, and sharing with this community.

[ThomazNeto]

Congrats on the Lovable role, Brahma, well deserved. Your POC posts are some of the few I actually go back to.

Since you asked what would make it useful, here's my take after a lot of POCs that never made it to production: the gap isn't ideas or tutorials, it's the path from demo to something a company will actually run. Three things I'd read every week:

Post-mortems, not just demos. Which POCs died, and why? Usually it's cost per query, latency, data quality, or security saying no. A short "what broke" section is worth more than the happy path.

The boring parts of the architecture. How the app authenticates: app service principal or user authorization, and what that means for Unity Catalog permissions. That decision alone changes the whole governance story of a Data + AI app, and very few write-ups cover it.
link doc

How you measured quality. Whether it's a Genie space, an agent, or a RAG app, show the eval set and the numbers before and after tuning. That's what convinces a steering committee, and it's what most community content leaves out.

If each project came with a repo deployable through bundles, so people could run it in their own workspace in ten minutes, that would set it apart from most of what's out there.

Count me in for contributing a builder story or two.

[Brahmareddy]

Thank you, @ThomazNeto. This is exactly the kind of input I was hoping to get.

I really like the focus on the demo-to-production gap. Post-mortems, authentication and governance decisions, cost/latency, evaluation results, and deployable reference projects are the areas that often get skipped in POC write-ups.

I am already thinking about a “what broke / what changed / what we learned” format for the builder stories, and your point about showing evaluation before and after tuning is especially valuable.

And yes, I would absolutely love to have you contribute a builder story or two. Great experiences like yours are exactly what can make this useful for the community.

Thanks again for sharing this.