A computer science doctoral student at the 91ĆĆ˝â°ć’s College of Natural Sciences and Mathematics is honored for his contributions to artificial intelligence (AI).
Rabimba Karanjai was recently named to the Google Cloud Research Innovators Program, where he is the only student in his cohort. He also received the Academic Research Award from the Sui Foundation for the third consecutive year, an achievement that previously only Cornell University had achieved.
Karanjai, who originally studied electrical engineering, said he was fascinated by computer science, saying coding allowed him more creativity. He turned his focus to AI reasoning and code generation, where he aims to make AI reliable and understandable.
Research to Recognition
Karanjai was nominated for the Google Cohort while collaborating with a researcher at Google. He is the only student in the group. “I became a cloud research innovator,” said Karanjai. “It opened up several research credits for me and allowed me to ask resources at Google for advice when studying my work.” In addition to the cohort, Karanjai is a Google Development Expert, who specializes in AI.
Karanjai’s exploration was also selected for the Sui Foundation’s Academic Research Award for a third straight year, an achievement only achieved by Cornell University.
Research for each grant works with the next. “We created an AI model which can take an existing smart contract in a defined programming language and translate a different one to ours,” said Karanjai. “In the third grant, we proposed a more secure aspect where we build a tool that can translate the accuracy of the code that you’re writing. It will suggest patches to make it work better.”
Motivation Beyond the Awards
Karanjai plans to keep his research in academia, with plans to become a professor. “These accolades boosted my confidence in what I’m doing,” said Karanjai. “It’s been awesome for me to give talks and steered me toward becoming a professor.”
He hopes to apply and extend his research to other fields, including biomedical and
legal domains.
AI might not always be correct, something Karanjai hopes fix. “It may fail two times
out of ten. Our research will showcase that point and will guarantee it when it works.”