Tanvi Aggarwal

Contact Info

Student Reseacher (Aug 2021 - Present)

Complex reasoning: Studying complex reasoning and generalization of compositional tasks in Natural Language Processing by prompting LLMs in zero-shot and few-shot settings.
Commonsense reasoning for answering why-questions: different aspects of incorporating external knowledge, including how to extract the most relevant information, how much knowledge to inject and in what form, and when to add it. Contributed towards building the model for extracting the most relevant relations.
Cross-modal knowledge distillation: Developed a cross-modal distillation strategy to improve over the performance of vision-only models for action anticipation.

Software Engineering Intern - StorageGRID (May 2022 - Dec 2022)

Supported metric conversion into different units for storage related metrics and white-labeling in the StorageGRID management tool using Angular.
This was a part of the interface overhaul to provide improved customer experience.

Software Engineer II - Revenue and Growth (Jul 2019 - Aug 2021)

Worked on multiple Go, Java, and Python backend microservices based on REST, gRPC framework with AWS cloud infrastructure.
Integrated AWS localstack cloud library in the local testing pipeline to reduce developer effort by 3 hours per feature deployment, and improve functional and unit test coverage by 35%.
Added middleware for error handling, system recovery and logging metrics to monitor services.
Led the design and deployment of several features on the discounting platform, including rate-limiting, tiered loyalty program (Swiggy Super) and a ranking system to show relevant offers based on consumer behavior patterns.
Worked on Named Entity Recognition for food items in code-mixed chat data using Deep learning, under the Applied Research team.

Technical Analyst Intern - Wealth Management (Jan 2019 - Jun 2019)

Implemented the server-side API for W3C’s Web Authentication specification for public key cryptography based secure-key authenticators using C#.
Integrated secure-key authentication with the existing multi-factor authentication system.

Software Engineering Intern - Office 365 (May 2018 - Jul 2018)

Created an on-boarding tool to migrate Data Loss Prevention Policies from compete platforms to Office 365 DLP using C# and XSLT.

Stony Brook University, New York (2021 - Present)
M.S. Computer Science
GPA - 4.0 / 4
Birla Institute of Technology and Science, Pilani (2015 - 2019)
B.E. (Hons) Computer Science
CGPA - 8.81 / 10

Text-Derived Knowledge Helps Vision: A Simple Cross-modal Distillation for Video-based Action Anticipation
Sayontan Ghosh, Tanvi Aggarwal, Minh Hoai, Niranjan Balasubramanian
Findings of European Chapter of the Association for Computational Linguistics (EACL 2023)
Using Commonsense Knowledge to Answer Why-Questions
Yash Lal, Niket Tandon, Tanvi Aggarwal, Horace Liu, Nathaniel Chambers, Raymond Mooney, Niranjan Balasubramanian
Main proceedings of Empirical Methods in Natural Language Processing (EMNLP 2022)
Softmax Coaxial Gaussian Track Code for Fingerprint Indexing
Geetika Arora, Tanvi Aggarwal, Kamlesh Tiwari
ACM India Joint International Conference on Data Science and Management of Data (CoDS-COMAD 2019)

Complex reasoning in NLP - Studying compositional generalization of NLP tasks by prompting LLMs in zero-shot and few-shot settings as part of MS thesis in Language Understanding and Reasoning Lab.
Approximate Pattern Matching using Seed-and-extend heuristic - Designed and implemented an optimized dynamic programming algorithm for approximate pattern matching in text strings using Python.
DNS resolver with added support for DNSSEC - Created a command line tool for a recursive DNS resolver to querry different record types with support for DNSSEC protocol, using dnspython library.
Mininet - Implemented static routes and dynamic routing (RIP) on a network topology configured using Mininet with AWS EC2.
Picolo: A Smart Assistant - Designed a system to improve the user interaction experience in online meetings that leverages OCR to extract text from images in real time. Implemented and evaluated a web-browser extension and desktop application interface.

Python, Go, C#, C/C++, Java, XSLT, HTML, Javascript, Ruby, SQL, Angular

Software development, Machine Learning, Databases, Algorithms

AWS (DynamoDB, EC2, DAX, Lambda), ElasticSearch, Redis, Git, Docker, Prometheus, Grafana, LocalStack, Jenkins, Snowflake, REST, gRPC

Pytorch, Tensorflow, Keras, NLTK, Pandas, NumPy, Transformers, LLMs, GPT-3, ChatGPT, Codex

Awarded the BITS Pilani Merit Scholarship for academic year 2015 - 2016 (awarded to top 3% of students).