Yuan Meng Email and Phone Number

I work on all things search: Query Understanding (QU), First-Pass Ranking (FPR), Second-Pass Ranking (SPR), AutoComplete, among other search stuff — sometimes out of business needs but often out of love :)"Enigmas of the universeCannot be known without a search"

As an international student, I taught classes to get funding — yet 4 years of teaching cultivated my love for explaining and providing resources for self-motivating mentees.Some classes I taught were in Python and some in R (which I sadly don't remember much about):- Python: CogSci 131 (Computational Models of Cognition), 2X- R: Psych 102 (Methods for Research in Psychological Sciences)- Psych 147 (Methods in Cognitive Development)- Psych C143 (Language Acquisition)- Psych 140 (Developmental Psychology)- Psych 124 (Human Evolution), 2X

As a child, I wondered about algorithms that underlie our thinking and how to teach a machine to think like us. Years later, I discovered that Cognitive Science aims to answer that very question, which led me to Cal.- My dissertation focused on active causal learning and feedback loops of racial disparities. My committee included Drs. Fei Xu, Alison Gopnik, Neil Bramley, and Mahesh Srinivasan.- I conducted over 10 studies using experiments, browser games, and computational models, publishing at seven top peer-reviewed conferences (CogSci 2017-2021, CDS 2018, SRCD 2021)."You insist that there is something a machine cannot do. If you will tell me precisely what it is that a machine cannot do, then I can always make a machine which will do just that!" — von Neumann (1948)

I did the "chat" with the then VP of ML @Bombora while hiking in the Blue Ridge Mountains, only to realize that was called a "behavior interview" later. I was lucky to get the job despite my lack of interview skills and work with a fantastic mentor who designed a fun project and afforded more trust and independence than I'd ever hope for:- Built end-to-end pipeline (PyTorch-BigGraph, PySpark) for embedding large-scale IP-URL graph with billions of edges and classifying IPs based on embeddings and other features- Scaled up pipeline on Google Cloud Platform; substantially enhanced classification precision of work-from-home IPs (0.47 to 0.98) and VPNs (0.63 to 0.99)