Akshita Bhagia
Senior Research Engineer at Allen Institute for AI (AI2)
Seattle, Washington, United States
Summary
🤩
Rockstar🎓
Top SchoolAkshita Bhagia is a Lead Research Engineer at AI2 with seven years of experience designing and productionizing NLP and ML systems, particularly in research-to-production workflows. She has deep hands-on experience with AllenNLP and allenai/allenlp-models—contributing fixes to training, distributed metrics, and model cards for core NLP tasks like QA and coreference. Her background blends academic research (UMass Amherst) on interpretability, dataset construction, and probabilistic annotation models with applied engineering across startups and finance, enabling her to bridge experimental research and robust engineering. Based in Seattle, she’s known for debugging subtle metric and logging issues in distributed settings—an underappreciated skill that improves model reliability in real-world deployments.
7 years of coding experience
5 years of employment as a software developer
Bachelor of Technology (B.Tech.), Information and Communication Technology, 8.95/10, Bachelor of Technology (B.Tech.), Information and Communication Technology, 8.95/10 at DAIICTMaster of Science - MS, Computer Science, 4.00/4.00, Master of Science - MS, Computer Science, 4.00/4.00 at University of Massachusetts Amherst
English, Hindi
Github Skills (18)
pytorch10
python10
model-driven10
machine-learning10
allennlp10
model-building10
deeplearning-ai10
deep-learning10
natural-language-processing10
nlp10
modeling10
metric10
model-driven-development10
testing9
develop9
Programming languages (6)
Github contributions (5)
allenai/allennlp-models
Jul 2020 - Jul 2022
Officially supported AllenNLP models
Role in this project:
ML Engineer
ML EngineerContributions:9 releases, 52 reviews, 56 commits in 1 year 11 months
Contributions summary:Akshita primarily contributed to the implementation of model cards, including the addition of a `ModelCard` class and related information for several models. They worked on updating existing metrics and integrating distributed metric functionalities, specifically in the context of question answering and coreference resolution tasks. Furthermore, they prepared for software releases and addressed formatting concerns.
nlppytorchldabertword-embeddings
allenai/allennlp
Jul 2020 - Jun 2022
An open-source NLP research library, built on PyTorch.
Role in this project:ML Engineer
ML EngineerContributions:9 releases, 196 reviews, 73 commits in 1 year 11 months
Contributions summary:Akshita made several commits related to improving the model's training and evaluation process. They addressed issues with logging, specifically related to regularization loss, and ensured the correct calculation of metrics. Additionally, they fixed a bug in the text classification predictor related to passing tokenized inputs. Furthermore, they implemented and debugged metrics used in distributed training environments, covering a variety of evaluation measures.
nlppytorchtransformerspythonbert
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.