Summary
Xide Xia is a research-driven engineer with 11 years of experience building state-of-the-art machine learning and computer vision systems, most recently contributing as a core member of Llama3 at Meta before moving to a stealth startup focused on "building something cool." He holds advanced degrees from Harvard and Boston University and has a strong publication record with papers at CVPR, ICCV, NeurIPS, ECCV and WACV, reflecting deep expertise in retrieval, hashing, metric learning, and visual representation learning. His background spans academic research, industry R&D at Google and Meta, and applied projects like real-time photorealistic style transfer and traffic-sign embedding that delivered measurable performance gains. Xide combines rigorous empirical methods from his PhD and teaching experience in stochastic optimization with practical product impact—evident in work that improved F1 scores and enabled deployment-ready models. Based in Menlo Park, he brings both the curiosity of an academic and the execution discipline of a senior research scientist, now channeling that mix into early-stage product development. An under-the-radar strength is his cross-domain fluency—from bioinformatics and astronomy-active learning to large-scale language model training—making him adept at translating research advances into novel applications.
11 years of coding experience
10 years of employment as a software developer
Master of Engineering - MEng Computer Science, Master of Engineering - MEng Computer Science at Harvard University
Doctor of Philosophy - PhD Computer Science, Doctor of Philosophy - PhD Computer Science at Boston University