Phoenix Wilson
/fiːnɪks wɪlsᵊn/[fiːnɪks wɪlsᵊn]phoenixw@berkeley.edu
/fiːnɪks wɪlsᵊn/[fiːnɪks wɪlsᵊn]Phoenix Wilson is a fourth-year EECS + Math transfer student at UC Berkeley interested in interpretability and formalisms of ML.
As an MLE intern at Esperanto Technologies, he utilized synthetic instruction tuning to train an LLM to perform classification on AI generated artwork. Subsequently, he conducted a literature review on Kolmogorov-Arnold Networks (KANs), a robust alternative to conventional MLP architecture.
At Foothill College, his work with the college's engineering club focused on researching and developing computer vision applications for robotics.
At EOPS and Pass the Torch, he served at-need community college students, providing one-on-one personalized tutoring and interactive lectures in Calculus, Statistics, Java, and Physics.
At Berkeley, he has served as a course tutor for EECS 16A as well as a junior mentor for Computer Science Mentors. Currently, he is an education consultant for Machine Learning at Berkeley, where he has completed a semester-long survey course on classical and modern-day ML.
Après cela, il y aura, j'espère, des gens qui trouveront leur profit à déchiffrer tout ce gâchis — Évariste Galois, 1832
Later there will be, I hope, some people who will find it to their advantage to decipher all this mess.