Rohith Pesala
Chief Technology Officer at Savante
Redmond, Washington, United States
Summary
🤩
Rockstar🎓
Top SchoolRohith Pesala is a technology leader and machine learning specialist with nine years of experience, currently serving as CTO in Redmond after leadership roles at Motional and applied science work at Microsoft. He bridges research and production, shipping computer vision and deep learning systems—from DETR-based table extraction contributions in a prominent Microsoft repo to scalable image segmentation pipelines at DeepMap. His background includes academic research in reinforcement learning and meta-MDPs as well as NLP work on dependency parsing, giving him a rare mix of principled research and practical engineering. Known for evolving model architectures, loss functions, and deployment/testing pipelines, he excels at turning novel ML research into robust products.
9 years of coding experience
6 years of employment as a software developer
Master’s Degree Computer Science, Master’s Degree Computer Science at University of Massachusetts AmherstBachelor of Technology (B.Tech.) Electrical Engineering, Bachelor of Technology (B.Tech.) Electrical Engineering at Indian Institute of Technology, Mandi
English, Telugu, Hindi, French, German
Github Skills (13)
data-preprocessing10
computer-vision10
data-loading10
machine-learning10
pytorch10
dataprep10
trainings10
preprocessing10
dtr10
python10
load-data10
modeling10
preprocess10
Programming languages (1)
Github contributions (3)
microsoft/table-transformer
Jun 2021 - Mar 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:29 commits, 9 PRs, 10 pushes in 9 months
Contributions summary:Rohith contributed to the core codebase of a deep learning model for table extraction. They added initial model components, including the DETR model itself. Their changes involved modifications to model architecture, loss functions, and data loading, particularly related to table detection and structure recognition, as evidenced by the inclusion of DETR related code. Furthermore, the user updated scripts and configurations for a testing pipeline, showcasing model deployment and metric computation capabilities.
deep-learningtable-detectiontable-extractiontable-structure-recognitiontable-functional-analysis
hari170894/PokerProject
Oct 2016 - Oct 2016
Contributions:1 push in 1 day
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.