Sep 2018 – Nov 2021 · 3 yrs 2 mos · San Francisco Bay Area

• I work on Natural Language Understanding.

Feb 2018 – Aug 2018 · 6 mos · Baltimore, Maryland Area

• Simulated High-Performance Computing (HPC) systems with natural and artificially injected faults using The Structural Simulation Toolkit (SST).
• Created a framework using python to perform node-based and task-based reliability analysis on logs generated by simulated HPC systems. This analysis is independent of Network Structure.
• Built a Support Vector Machine based classifier to identify artificial fault injection. Weibull and Log-Normal lifetime models were used to parameterize the reliability curves.

Sep 2017 – Jan 2018 · 4 mos · Sunnyvale

• Simulated human annotators using Bayesian modeling to create synthetic annotated data for Speaker Identification (SID) system.
• Used Unsupervised Label Refinement (ULR) methods (like Dawid Skene) and showed that these methods work better than Majority Voting for SID annotation.
• Evaluated human annotator's False Acceptance Rate (FAR) and False Recognition Rate (FRR) for speaker identification by created ground truth data of varying difficulty.
• Showed that the current annotation process was unacceptable even if we use ULR on labels from multiple annotators to reduce errors.
• Showed that metadata of test utterance and enrolled utterance did not have enough signal to judge annotation difficulty.
• Created training and testing data to evaluate domain classifier.

Jun 2017 – Jul 2017 · 1 mo · Baltimore, Maryland

• Created training data for localization and detection of Zika virus-carrying mosquito species for a custom built trap.
• Used Faster Region-based Convolutional Neural Networks (R-CNN) to localize the mosquito species and Residual Neural Network (ResNet) to detect mosquito species with 81% accuracy.
• The results of this work helped in getting funding for this project.

Nov 2015 – May 2016 · 6 mos · Lille Area, France

• Integrated clinical and -omics data of Acute Myeloid Leukemia (AML) from The Cancer Genome Atlas (TCGA) Multimodal Representation Learning.
• Used Regularized Generalized Canonical Correlation Analysis (RGCCA) and Sparse Generalized Canonical Correlation Analysis (SGCCA) for creating multimodal representation and selecting important genes.
• Suggested design matrices for RGCCA by using graphical methods from MixOmics package and by considering biological results already provided by The Cancer Genome Atlas (TCGA) consortium.
• Reproduced results of Lasso-based models of data integration and a paper on Survival Analysis.

Jun 2014 – Jul 2014 · 1 mo · Kazakhstan

• Windows Server 2012
• Project Mentor: Alexandr Chsherbov
• Installed server 2012 and maintained 50 clients.
• Activated services like authentication, mailing, active directory service, internet access, remote client management and group policy management.