Hello World, I'm Sara.

I am a music technologist, machine learning engineer and musician working to pioneer new ways of creating and performing music with AI. I currently work as a Senior Machine Learning Engineer at Neural Magic by day, developing open source software for high-performance LLM inference. By night I make electronic music with my guitar and laptop, using a combination of Tidal Cycles live coding and electric guitar looping. I received my Masters of Science in Sound and Music Computing at Queen Mary University of London, where I was funded through a US-UK Fulbright grant to pursue music AI research in the UK. My master's thesis focused on developing a steerable Transformer-XL model capable of generating loopable musical phrases for use in performances. It was published at EvoMUSART 2023 where it received the Outstanding Student Award. I continue to incorporate generative AI into my performance practice, focusing on human-AI improvisation and timbre synthesis. I have given talks on music AI to audiences of both musicians and researchers at venues such as Music Hackspace, ISMIR, Hackaday Superconference and XFest.

Prior to my time in London, I spent three years in Boston working as a machine learning engineer at Bose and performing as a freelance classical guitarist for weddings and other events. While at Bose, I worked in the applied research group in the Health division, developing prototypes that resulted in two published patents and two pending. I worked on deep learning models for speech enhancement, writing custom kernels to optimize quantized models to run live on a hearing aid micro-controller. I also led a research project developing generative audio algorithms that adapt to biofeedback signals to induce sleep using soothing music.

I hold an interdisciplinary bachelor's degree in Music Technology and Computer Science from Carnegie Mellon University. While in Pittsburgh, I performed as a guitarist with chamber group pairings ranging from renaissance vocalists, to string quartets, to banjo players. My senior capstone project, “Creating with the Machine,” was a set of compositions combining algorithmic and traditional methods of composition into live performances to explore how interactive generative algorithms can influence creativity in musical improvisation. “Creating with the Machine” was premiered by the Carnegie Mellon Exploded Ensemble, and was awarded the Henry Armero Memorial Award for Inclusive Creativity. I also presented the project at the Hackaday Superconference in LA.

I pride myself on having performed not only in concert halls, but also in an underground limestone mine and a giant inflatable dream temple. As guitarist, violist and composer for the Exploded Ensemble in Pittsburgh, I collaborated with other artists and musicians to put on a literal underground electronic music festival at Bradys Bend Limestone Mine. We also performed an 8 hour overnight concert of ambient music, where audience members were invited to drift in and out of sleep in an artist-created inflatable environment to experience the subconscious effects of the music. Needless to say, I embrace abnormal and experimental performances. I'm inspired by musical ideas that are new, different and weird, and am dedicated to making these types of performances accessible and enjoyable for all audiences.