Brandon Smock
Senior II Applied Scientist, Document Intelligence at Kensho Technologies
Seattle, Washington, United States
Summary
🤩
Rockstar🎓
Top SchoolBrandon Smock is a Senior II Applied Scientist specializing in document intelligence with 11 years of experience bridging research and product-focused machine learning across academia and industry. He holds a Ph.D. in Computer Engineering from the University of Florida and has driven applied research at Microsoft and Kensho, delivering solutions in document understanding, OCR, object detection, and multi-modal vision-language models. Brandon contributes to open-source evaluation work—most notably improving metrics and reporting for Microsoft’s Table Transformer repository—bringing rigor to model assessment and reproducible research. He balances big-picture ideas with hands-on engineering, favoring simple, elegant solutions for complex sequence and transformer-based problems. Based in Seattle, he also applies ML to diverse domains from remote sensing and landmine detection to molecule generation, signaling a practical curiosity that finds common structure across varied problem spaces.
11 years of coding experience
20 years of employment as a software developer
Doctor of Philosophy (Ph.D.) Computer Engineering, Doctor of Philosophy (Ph.D.) Computer Engineering at University of Florida
English
Github Skills (7)
machine-learning10
python10
evaluation10
metric10
deep-learning9
pytorch9
computer-vision8
Programming languages (1)
Github contributions (5)
microsoft/table-transformer
May 2021 - Nov 2022
Table Transformer (TATR) is a deep learning model for extracting tables from unstructured documents (PDFs and images). This is also the official repository for the PubTables-1M dataset and GriTS evaluation metric.
Role in this project:
ML Engineer
ML EngineerContributions:1 release, 103 commits, 24 PRs in 1 year 6 months
Contributions summary:Brandon primarily contributed to the evaluation and metric reporting aspects of the `Table Transformer` project. They added and refined metrics, including GriTS and DAR, for assessing the performance of table extraction models. Their work involved modifying the `grits.py` and `eval.py` files to enhance the readability and standardization of the metric outputs, as well as fixing issues in the reporting. They also worked on simplifying and optimizing the core evaluation function.
extractiondatasetevaluationmachine-learningmodel-training
bsmock/k-shortest-paths
Nov 2015 - Nov 2017
Contributions:1 release, 18 commits, 2 PRs in 2 years
k-shortest-pathsnodesyenpathsgraph
Find and Hire Top DevelopersWe’ve analyzed the programming source code of over 60 million software developers on GitHub and scored them by 50,000 skills. Sign-up on Prog,AI to search for software developers.