Design Once, Deploy at Scale: Template-Driven ML Development for Large Model Ecosystems

Jiang Liu, John Martabano Landy, Yao Xuan, Swamy Muddu, Nhat Le, Munaf Sahaf, Luc Kien Hang, Rupinder Khandpour, Kevin De Angeli, Chang Yang, Shouyuan Chen, Shiblee Sadik, Anirudh Agrawal, Djordje Gligorijevic, Jingzheng Qin, Peggy Yao, Alireza Vahdatpour · 2026

Read Design Once, Deploy at Scale: Template-Driven ML Development for Large Model Ecosystems by Jiang Liu, John Martabano Landy, Yao Xuan, Swamy Muddu, Nhat Le, Munaf Sahaf, Luc Kien Hang, Rupinder Khandpour, Kevin De Angeli, Chang Yang, Shouyuan Chen, Shiblee Sadik, Anirudh Agrawal, Djordje Gligorijevic, Jingzheng Qin, Peggy Yao, Alireza Vahdatpour (2026) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2603-24963.txt.

Description

Liu and co-authors study how one shared template can build many ad ranking models at a large platform. Each model used to be tuned on its own, which slowed new ideas. Their Standard Model Template builds models from shared parts that fit many goals and data types. Over four release cycles at Meta, the template cut cross-entropy by 0.63 percent at the same serving cost. It also cut per-model build time by 92 percent and sped up technique rollout more than sixfold.

Catalog record

CREATORJiang Liu, John Martabano Landy, Yao Xuan, Swamy Muddu, Nhat Le, Munaf Sahaf, Luc Kien Hang, Rupinder Khandpour, Kevin De Angeli, Chang Yang, Shouyuan Chen, Shiblee Sadik, Anirudh Agrawal, Djordje Gligorijevic, Jingzheng Qin, Peggy Yao, Alireza Vahdatpour
YEAR2026
COLLECTIONPAPERS
FORMATTXT
FILE SIZE38695
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2603.24963

Browse the full library · About the Tower of Babel