▲ About:
-
In May 2022, I earned my M.S. in Electrical and Computer Engineering at University of Texas at Austin, with a
focus on machine learning.
Previously, I worked as an Applied Scientist Intern at Amazon Search, a Machine Learning Engineer at ASAPP and as a Fulbright Scholar focused on NLP for low-resource Indian languages.
▼ Work Experience:
-
Applied Scientist Intern | Jun 2021 - Aug 2021
-
Explored the feasibility of using relational graph convolutional networks (RGCNs) for product and query representation learning.
-
Demonstrated that pseudo-labeling and knowledge distillation with a graph-based teacher model can be used to improve downstream classifiers with limited data.
-
-
-
Machine Learning Engineer | Feb 2019 - Aug 2020
-
One of two engineers responsible for the type-completion service of a customer interaction platform, which received 100+ requests per second.
-
Automated majority of model lifecycle, from dataset creation to model retraining, evaluation, and versioning.
-
A/B tested new models and analyzed test results to understand and communicate impact. Automated the setup and analysis of routine A/B tests (e.g., A/B testing models retrained on new data).
-
Led the research and development for a personalized type-completion model to increase the efficiency of customer service agents.
▼ Research Experience:
-
U.S. Fulbright Scholar | Fall 2017 - Spring 2018
-
Hyderabad, India
-
Kelsey Ball, Dan Garrette. “Part-of-speech tagging for code-switched and transliterated text without explicit language identification.” In Proc. of EMNLP, 2018.
-
My host institution, IIIT-H, published an article about my experience.
-
Research Assistant | Fall 2014 - Fall 2015
-
University of Texas at Austin
-
Developed statistical models to analyze code-switching in bilingual speakers. Presented at two conferences on language contact and transfer.
-
The Undergraduate Research department at UT wrote a profile on my work.
▼ Graduate Course Projects:
-
Course: Statistical Machine Learning and Optimization (SDS 384)
- Final Project: Unsupervised Contrastive Representation Learning: A Survey (pdf) (slides)
-
Course: Natural Language Processing (CS 388)
- Projects: Named Entity Recognition (pdf), Sentiment Analysis (pdf), Semantic Parsing (pdf)
- Final Project: A Broader Evaluation of Multilingual BERT on Code-Switched Data (pdf)
-
Course: Deep Learning (CS 395T)
- Final Project: Window-Based Activation Functions (pdf)
-
Course: Online Learning (ECE 381V)
- Final Project: Preference-Based Multi-Armed Bandits (slides)
-
Course: Information Theory (ECE 381K)
- Final Project: An Information-Theoretic Analysis of Thompson Sampling (slides)
-
Course: Statistical Learning Theory (EE 381V)
- Final Project: An Overview of the Equivalence Between Online Learning and Differentially-Private PAC Learning (pdf)
▼ Teaching:
- I have worked as a teaching assistant for two undergraduate courses and one study-abroad engineering course. I've also received a Teaching Award from the Cockrell School of Engineering.
-
Course Professor Semester EE319K Dr. Ramesh Yerraballi Spring 2014 EE319K Dr. Mohit Tiwari Spring 2016 EE306 Dr. Ramesh Yerraballi Fall 2016 Embedded Systems and Industry in India Dr. Ramesh Yerraballi Summer 2016
▼ Service:
-
Texas 4000 for Cancer
-
I co-led a team of 25 students cycling over 4,000 miles from Austin, TX to Anchorage, AK. We raised over $800,000 for cancer research and support services, making us the highest fundraising team in the organization's history.
-
Learn more about or see footage of the journey.
-
-
Longhorn Stream Team
-
I helped form a citizen-science initiative at UT. We paddle rivers and collect water quality data to promote water conservation.
-
We also volunteer in the community by teaching environmental science to kids, and some of our members are involved in active research. See more of what we do.
-