Varad Pimpalkhute
Research Scientist at MBZUAI (Mohamed bin Zayed University of Artificial Intelligence)
California, United States
Summary
👤
Senior🎓
Top SchoolVarad Pimpalkhute is a research scientist focused on applying large language models, retrieval-augmented generation, and meta-learning to real-world problems, with eight years of experience spanning industry and academia. He developed a novel LLM-informed query encoder at Amazon (SoftQE) that yielded measurable OOD gains without added latency and led to a paper accepted at ECIR'24. His work ranges from failure-mode analysis of RAG pipelines and automated evaluation for multi-turn dialogue to RL-driven schema generation for moral reasoning, reflecting a blend of systems engineering and rigorous empirical research. At MBZUAI he now advances foundational models with asynchronous RL training and reasoning, building on prior applied research roles at Articul8 AI and AI2. An active open-source contributor, he improved usability and docs for the prominent learn2learn meta-learning library, demonstrating attention to reproducibility and developer experience. Colleagues note his uncommon combination of production-minded implementation and deep theoretical curiosity, evidenced by a perfect MS GPA and a track record of cross-disciplinary internships.
8 years of coding experience
3 years of employment as a software developer
Master of Science - MS, Computer Science, 4.0/4.0, Master of Science - MS, Computer Science, 4.0/4.0 at University of Massachusetts AmherstDiploma, Diploma at Somalwar Academy Education SocietyBachelor of Technology - BTech, Electrical, Electronics and Communications Engineering, Bachelor of Technology - BTech, Electrical, Electronics and Communications Engineering at Indian Institute of Information Technology Nagpur
English, Hindi, Marathi
Stackoverflow
648reputation
90kreached
5answers
3questions
Github Skills (18)
pytorch10
python10
meta-learning10
machine-learning10
few-shot-learning10
mask-rcnn9
faster-rcnn9
documentation8
unit-testing7
pip6
shell6
anaconda6
scikit-learn6
virtualenv6
linux6
Programming languages (12)
Github contributions (5)
learnables/learn2learn
Oct 2021 - Mar 2022
A PyTorch Library for Meta-learning Research
Role in this project:
ML Engineer
ML EngineerContributions:1 review, 8 commits, 5 PRs in 5 months
Contributions summary:Varad primarily contributed to the `learn2learn` library by fixing bugs, improving code examples, and updating documentation related to meta-learning research. Their work involved addressing errors in existing implementations, such as the ANIL example and the RandomClassRotation transform. Additionally, the user added tutorials and updated documentation generation processes. The commits show an active role in improving the usability and functionality of the library.
pytorchmeta-learningmetalearningmeta-optimizationdeep-learning
iiit-nagpur/website-PC
Sep 2018 - Oct 2018
Website for programming community at IIITN(Under Development)
Contributions:22 commits, 9 PRs, 17 pushes in 19 days
under-developmentiiitn
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.