Aashish Kolluri
Crystal Center. KiSP Lab. School Of Computing. National University Of Singapore.

AS06-04-25
Singapore, 117416
Bio
Hi! I am a PhD student at National University Of Singapore. I am doing my thesis with Prateek Saxena. I design scalable and trustworthy machine learning algorithms for learning on structured data such as graphs. I work on research problems that are motivated by practical relevance and at the same time have theoretical guarantees. In general, I am a systems security person and I am inclined to analyze the security of any newly introduced frameworks/ideas across all subfields of computer science. Attacks are interesting but defenses with guarantees are beautiful to me.
I have experience in program analysis, synthesis, debugging and security of blockchains.
History
In 2017, I graduated from IIT Kanpur where I was fortunate to work with Indranil Saha. I visited Prateek in 2017 for an internship before joining him as an advisee.
Work
I was fortunate to work with two exciting companies, Flipkart India and Aqilliz. Flipkart India (now owned by Walmart) gave me an opportunity to work on their anomaly detection engine. Aqilliz is an exciting startup from Singapore where I had a chance to demonstrate the impact of applying Differential Privacy protection on the utility of widely-used queries like recommendations.
Other interests
My interests range widely from Computer Science to Psychology, Astronomy, Music, Cricket and Animals. I love singing and relish performing almost anything on a stage(of course after practicing….). I will probably stay in Singapore for at least another three years, so, If you are here then feel free to drop by and say hoi.
news
Mar 1, 2023 | [New!] Please find our work on scalable neural network training on distributed graphs on Arxiv here. |
---|---|
Jan 1, 2023 | [New!] Our paper on user-customizable automatic transpilation, DuoGlot, from Python to Javascript has been accepted to OOPSLA’23. |
Aug 28, 2022 | Our paper on new graph neural network architetures, LPGNet, that provide privacy for edges got accepted to CCS’22. |
Jul 21, 2021 | Our paper on private hierarchical clustering in federated networks got accepted to CCS’21. |
Jun 21, 2021 | Our paper on fault localization using a single exploit got the best paper award at AsiaCCS’21! |