Cristian Dinu
  • Writing
  • Tablete
  • Talks
  • About

Cristian Dinu

Research software, energy, AI and society. Writing and work by Cristian Dinu.

Engineering.
Uncertainty.
And being human.

I’m a software engineer working across energy, data and research. I test how AI systems and agents behave under pressure, and build data infrastructure for UK household energy studies. This site is where I think in writing: essays on technology and the people around it, and shorter, more personal pieces.

A little more about me

Cristian Dinu
Engineer, researcher,
occasionally wrong.

The public notebook

Recent writing

All writing

AI Needs Assurance

Software that calls LLMs at runtime is becoming common infrastructure, and infrastructure needs assurance. AI evaluations are one of its building blocks.

19 Oct 2026

What has been keeping me busy, September 2026

AI is at the centre of what I’ve been doing, unsurprisingly. But testing and evaluating LLMs is only half the story — the other half is why I still insist on…

25 Sep 2026

From Bitcoin to AI

I’ve made tech predictions that soared and others that crashed: from laughing at Bitcoin to betting on Linux. Turns out I’m sometimes good at spotting solid…

29 Jul 2025

Dashed — Writing Like a Machine

I like dashes of all kinds: the hyphen, the syllable-breaking dash, the dialogue dash. I like diacritics, I like commas. Whenever I can, I use non-breaking…
23 Jul 2025

Power Shift: From Passive Consumers to Grid Players

Energy flexibility means the ability to adjust power generation and demand, so that they match exactly. It’s going to be a big topic in the following years…

22 Jul 2025
No matching items

Shorter, more personal

Tablete

All tablete

Comrade Engineer Dinu

An archive of old newspapers, a forgotten genre, and the engineer I might have been at the printing works.

25 Oct 2026

The Voucher

Who deserves a few hundred more years of healthy life? Not me. Teachers, mostly.

24 Oct 2026
No matching items

Open questions

Thinking through, not finished with.

Lately I keep coming back to the same question: how do you evaluate and test software built on non-deterministic components, particularly LLMs and agents?

  • How do you establish what “good enough” means?
  • How do you investigate failures, and tell whether a change has actually improved the system?
  • What can simulations tell you about the real world?

More about my current interests

Ideas in practice

Work & talks

Energy research

Data systems for SERL & EDOL

My work focuses on designing and building data infrastructure that gives UK researchers access to detailed energy datasets while maintaining strong privacy protections.

About the work · EDOL update

RSECon25 · Talk & demonstration

The End of Front-End

Delivering software via LLMs and the Model Context Protocol.

Talk & materials · Demo code

© 2026 Cristian Dinu
Views here are my own.
Last updated: 27 Oct 2026, 11:40 UTC

 

RSS: writing · tablete
Powered by Quarto