daita@system:~$ cat ./about.md
# About Daita
A small R&D shop working on hard problems in applied AI, data engineering, and cloud software.
## What we do
We pick research questions where the gap between a paper and a working system is the actual difficulty, then close that gap. Agent harnesses, evaluation, retrieval over messy data, lakehouse pipelines, distributed systems, platform engineering.
We publish what we learn. Open-source benchmarks, reference implementations, and technical writeups. The blog is the primary artefact.
We collaborate with teams who need a research partner more than a vendor. Short, scoped engagements. Co-authored output where it makes sense.
## How we work
* Research, then ship
A prototype that runs on real data beats a slide that describes one. We measure, then we write.
* Open by default
Code, benchmarks, and writeups go public unless a collaboration says otherwise. The work compounds.
* Small on purpose
Two co-founders plus a tight network. Senior attention on every problem; no layered hand-offs.
* Honest about limits
We tell you when a problem is solved, when it is not, and when the answer is "do not build this".
* Reproducible
Pinned environments, seeded runs, golden vectors. If we cannot reproduce a number, we do not publish it.
* Curious by trade
We pick problems we want to understand more deeply. That bias shows up in the depth of the output.
## The team
Two co-founders with deep experience across backend development, data engineering, cloud infrastructure, and AI. Small by design. Every research question gets senior attention.
## Collaborate
If you have a research-shaped problem and want a partner who will take it seriously, drop us a line. We will tell you honestly whether we can help.
## Partners
We team up with trusted companies when a project needs it. Real collaborations with people we have worked with and vouch for.
daita@system:~$ _