I'm a bioinformatics scientist with expertise ranging from analyzing multi-omic datasets to implementing high-performance sequence alignment algorithms. I currently work at PacBio as Senior Bioinformatics Software Engineer, developing on-instrument algorithms that turn raw sensor data into highly accurate DNA reads.
In language models, the order of tokens is of critical importance. In this post I'll explain RoPE: a technique to ensure the pairwise token attention weights depend on the relative position to each other instead of absolute position in the sequence, improving model generalization.
K-mers are frequently used in bioinformatics to profile and compare genomes. This article describes a technique for compactly representing long k-mers with locality-sensitive hashing.
The Netherlands is very flat country, with most its area below sea level. So how flat is it really, and how to we keep the water at bay? Let's visualize the official "height" dataset of the Netherlands to find out!
Through some clever mathematical definitions, we can use OpenGL "point" primitives to draw any kind of vector shapes. This article explains how that's done and how it applies to my Google Summer of Code 2015 project.
Servos are useful small motors with applications in robotics, RC cars, and others. Here, we describe how to use an ATTiny44 to control its position.
