Anders Dalskov

me

About me

Hello, I am a cryptographic engineer / cryptographer from the great small country of Denmark. This webiste is a run-of-the-mill static personal webpage whose primary purpose is to provide me with a place where I can collect things relevant to me—such as information about me, my publications, and my thoughts (the ones I feel are worth sharing anyway).

Picture on the left is me. It's a bit old, but my looks haven't changed all that much since it was taken. I'm nicer than I look :-)

Interests

A particular area of cryptography for which I have a great deal of fondness, is Secure Multi-Party Computation. And one of the aims of this website is to write about this frankly neat technology.

Secure Multi-Party Computation (or MPC as its usually abbreviated) is a cryptographic tool that allows multiple parties to perform a joint computation without having to reveal their individual inputs to each other. This is most often done using some form of linear secret-sharing. Part of what makes MPC a fun technology to work with, is its flexibility. If the threat model assumes a very powerful adversary corrupting all but one party, then we get a fairly in-efficient protocol. On the other hand, if the threat model assumes an honest majority, then the protocol can be made much more efficient and may even provide additional useful features, such as being able to identify misbehaving parties, or even guaranteed output.

My current day-job is doing cryptography related stuff at Partisia, a Danish start-up/scale-up. Partisia is perhaps best known in the cryptographic community for helping Danish sugar-beet farmers trade sugar-beets using MPC. Much of what I do is related to (or semi-related to) MPC. Before that, I did a PhD at Aarhus University at the cryptography group there. Although my PhD was supposed to be more cyber-security oriented (partly initiated by my master thesis on insecure end-to-end encrypted cloud storage), the topic ended up being MPC focused, and specifically the use of honest-majority MPC for secure machine learning.

The webiste itself is built using emacs and org-mode, and hosted on the tiniest VPS that money can buy from ovh.

Contact


CC BY-SA