I am currently working as a Data Scientist in the Microsoft Security Research team. My work involves working on both endpoint protection (covering social engineering attacks such as phishing/tech scams) as well as end point detection and response (post-breach behavioral detections). My work is around anomaly detection, learning on graphs and multimodal learning.
I was previously a Master’s student at the University of Rochester, where I worked with Prof. Ralf Haefner on Uncertanity estimation using probabilistic Deep Learning. In summer 2020 I worked at Amazon Lab126, Sunnyvale California as an applied scientist in the Alexa AI team on pre-training graph neural networks for NLP tasks. Prior to that I worked at NeuralSpcae as a research Engineer. I did my undergraduate thesis at TU Kaiserslautern, Germany under Prof. Marius Kloft on Interpretability in Multimodal Deep Learning. I have briefly visited Prof. Marcus Liwicki’s lab at Lulea Tecchnical University, Sweden where I worked on bot detection on tweets. I have also worked on arithmetic word problem solving (IJCNLP 2017) and probabilistic activation functions(NeurIPS 2020). I have also served as a reviewer for ECML and NeurIPS 2020.
In my free time I love to play Badminton, write and hit the gym!
|Feb 1, 2021||Talk on Multimodal learning and graph neural networks in NLP at the VITA group @TU Austin|
|Jan 1, 2021||Graduated and joined MS Security research|
|Dec 2, 2020||ProbAct: A Probabilistic Activation Function for Deep Neural Networks accepted in NeurIPS OPT ML workshop.|