Systems for Artificial Intelligence Lab(SAIL)• Systems research for Deep Learning supervised by Dr. Alexey Tumanov• Design supernets for natural language processing models such as BERT for efficient model training and deployment in hardware-constrained settings.• Develop a new training and inference system by building on top of Hugging Face library in PyTorch.

• Social and Language Technologies (SALT) Lab• Natural Language Processing Research supervised by Prof. Diyi Yang• Research on Humor Classification and Generation [Fall 2020 - Spring 2021] • Developed humor detection and generation models utilizing various natural language processing models, including BERT, BART, and GPT2, via frameworks such as Hugging Face. • Designed methods to evaluate machine generated humor both quantitatively and qualitatively.

Teaching assistant for CS2110 Computer Organization and Programming.The class covers an overview of the basic structure and operation of a modern computer. Main topics include digital circuits, machine language instructions, and C programming.

Worked as Teaching Assistant for Georgia Tech's CS 1332 Data Structures and Algorithms course. The class covered basic data structures including queue, stack, tree, and heap, and also algorithms related to graph and dynamic programming.

Alexa Organization• Developed a multimodal Alexa skill that provides an optimized interactive user experience via utilization of implicit contextual information.• Established a scalable webapp infrastructure incorporating AWS services: Lambda, DynamoDB, S3, CloudFront, etc.• Leveraged AWS Rekognition to utilize computer vision techniques in identifying the contextual data.

Data Science at Georgia Tech(DSGT) is Georgia Tech's premier student-run data science organization. We host Hacklytics, an annual data science conference and hackathon bringing together over 300 student technologists, aspiring data scientists, and future engineers from Georgia Tech and across the southeastern United States.I am a corporate outreach person for both DSGT and Hacklytics, being part of External Affairs and Operations team respectively. I reach out to potential sponsors well as plan workshops and other special events for the organization.

Software Platform Lab(SPL) at Seoul National University studies machine learning systems and algorithms, big data systems, and datacenter software. We develop tools for training and testing deep learning models in various settings. • Developed data refurbishing, a new approach in data augmentation to make Deep Learning model training faster.• Implemented Revamper, a data loading system realizing data refurbishing on top of PyTorch library with 2K+ lines of code. • Accelerated the training of computer vision models(VGG16, ResNet-18, MobileNet-V1, EfficientNet-B0) by 1.03× – 2.04× while maintaining comparable accuracy.

Collected data from surveys and created excel files for statistical analysis that would be provided to different companies