Aman Rangapur
Research Engineer at Ai2
Seattle, Washington, United States
Summary
🤩
Rockstar🎓
Top SchoolAman Rangapur is an LLM Solution Architect based in Seattle with seven years of experience building and optimizing language and vision models for real-world impact. Currently at the Allen Institute for AI, he focuses on model-level optimization for OLMo research and has contributed to the allenai/olmocr toolkit by restoring critical modeling components and improving vision backbone and attention configurations. His background spans cross-domain authorship attribution, fact-checking dataset creation, medical QA fine-tuning, and transformer-based rotated object detection and segmentation, showing a rare blend of multimodal expertise. Aman combines hands-on full-stack development, ML infrastructure ownership, and academic research, having moved from robot vision on tiny hardware to production-ready LLM solutions. He brings a practical, results-driven approach informed by graduate studies in computer science and a track record of fixing subtle modeling issues that unblock downstream research.
7 years of coding experience
6 years of employment as a software developer
Master of Science - MS Computer Science, Master of Science - MS Computer Science at Illinois Institute of TechnologyBachelor's degree Computer Science, Bachelor's degree Computer Science at Vellore Institute of Technology
Spanish, English
Stackoverflow
1reputation
0reached
0answers
0questions
Github Skills (9)
attention-mechanism10
develop10
python10
vision-transformer10
configuration-management9
pytorch9
machine-learning8
data-structures8
data-structure8
Programming languages (7)
Github contributions (5)
allenai/olmocr
Feb 2025 - Sep 2025
Toolkit for linearizing PDFs for LLM datasets/training
Role in this project:
Full-stack Developer
Full-stack DeveloperContributions:2 reviews, 17 PRs, 86 pushes in 7 months
Contributions summary:Aman resolved code style issues and updated the project's documentation. They restored a crucial modeling file, indicating a potential fix or re-integration of a critical component. Furthermore, the user made changes within the modeling code, updating configurations and code. This work involved modifications to vision-related backbone configurations and attention mechanisms.
llm
IIT-DM/Fin-Fact
Aug 2023 - Sep 2024
A Benchmark Dataset for Multimodal Scientific Fact Checking
Contributions:1 review, 3 PRs, 90 pushes in 1 year 1 month
multimodaldatasetdatasetsfact-checkfact-checking
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.