Kevin’s Homepage

I lead a research group. I’m using computational tools and machine learning to capture the tacit and fuzzy dimensions of chemistry and materials science – to design materials that work in the real world.

Check out the lab website lamalab.org – we’re hiring!

Contact

Errors and Corrections

I offer €5 for minor errors, €15 for moderate errors, and €40 for major errors in my work. While it’s difficult to define these categories in advance, I aim to be fair in my assessments.

Very minor errors (typos, grammatical mistakes) earn my gratitude.

If you prefer not to receive a reward, I’ll donate the equivalent amount to GiveWell.

Contact me via email if you want a reward, or use the anonymous feedback form if you prefer anonymity.

Consulting

I provide pro-bono consultations (2 hours/month) for non-profits and organizations working for the public good who think they can benefit from my experience in ML, chemistry, and data management.

For students interested in mentoring, there is a program my lab is starting at lamalab.org/mentoring - please reach out via that one.

I’m also open to discussing consulting and advice for for-profit organizations for a fee.

Contact me via email for either.

Notes

A collection of unfinished thoughts and ideas.

Subscribe to get notified of new posts

RSS Feed: Add this RSS feed to your favorite RSS reader

Email Updates: Subscribe at Blogtrottr using the RSS feed URL: https://kjablonka.com/index.xml

Date	Title
1/27/26	Deep Networks as Elastic Origami
12/29/25	The Geometry of Not Enough Data
10/31/25	Why I Start Things Uncomfortably Early
10/14/25	On Impactful Research
10/1/25	Lessons Learned from Writing My First ERC Proposal
8/5/25	All of AI is Context Engineering
6/1/25	The Epistemic Risks of AI-Only Science
4/25/25	The Optimal Amount of Wasted Research Funding Is Non-Zero
4/23/25	Autoencoders as Digital Archaeologists for Spectroscopic Data
4/5/25	Notes on My Peer Review Process: An Invitation to Compare Practices
2/5/25	Beyond the Era of Accidental Discovery
2/2/25	10 Reasons to Aim Higher. And Higher.
1/12/25	Dear Claude: Are We Getting Too Close?
1/9/25	Thinking aloud about the shape of scientific data
1/5/25	A wise bird
1/5/25	Trust Me, There’s a Method to This Madness
12/2/24	Take it easy, my friend
9/21/24	The Tail End
5/3/24	Performing basic analysis of molecules generated by ML models
5/2/24	Building a GPT that can generate molecules from scratch
5/2/24	Building an LLM agent from scratch
3/24/24	The ‘researcher’s stuff’
3/2/24	Language I want to be more mindful of
3/2/24	Multiple instances learning
2/23/24	Developing an intuition for backpropagation