Experience
Software Engineer, Level IV @ Google LLC
Dec 2022 → Present
New York
- Working on Keep, a notetaking editor in Google Workspace.
Software Engineer, Level IV @ Google India Pvt Ltd
Aug 2020 → Nov 2022
Bangalore
- Developing intelligent features for the Google Workspace Editors (Docs, Slides, Keep, etc) using my expertise on the products' client-side software, supporting tools and libraries, and natural language processing infrastructure.
- Using cutting-edge frontend tools like Web Assembly and Emscripten, and Google-internal technologies like j2CI, client-side cross-platform frameworks and build systems, to develop user-facing features such as spellcheck in encrypted documents for five languages and writing style suggestions for English text.
- Formulating technical designs for independent end-to-end problems, driving cross-team collaboration, upholding software reliability practices, technical-debt resolution and documentation, and proactively identifying areas of future work.
- Guiding junior engineers on programming and software design tasks to enable timely delivery of products to customers.
Software Engineering Intern @ Google India Pvt Ltd
May 2019 → July 2019
Bangalore
- Worked on the Editors client-side software infrastructure to develop a user interface with control options to undo or provide feedback on the correction and a logging framework, for the Google Docs text auto-correction feature.
Research Intern, Big Data Experience Labs @ Adobe Research
May 2018 → July 2018
Bangalore
- Developed a mobile application for Text to Scene Conversion in Augmented Reality, based on novel research techniques for prediction of three-dimensional object sizes and positions from textual features.
Master's Thesis, Language Technologies Institute @ Carnegie Mellon University
Sep 2019 → May 2020
Carnegie Mellon University
- Machinated a novel technique for paraphrase generation using the von Mises-Fisher (vMF) Loss on a transformer network, and showed that it produces superior paraphrases as compared to the log-likelihood model by employing bilingual data to induce zero-shot paraphrasing, guided by Prof. Yulia Tsvetkov.
Research Intern, Department of Computer Science and Automation @ Indian Institute of Science (IISc)
May 2017 → July 2017
Bangalore
- Developed a cognitive text parser that combines syntactic and semantic approaches, to process textual data into cognitive structural representations, to be used as a feature extractor for downstream NLP tasks, and demonstrated the correlation of the extracted cognitive features with semantic and syntactic text features, guided by Prof. Veni Madhavan.
Natural Language Processing - Course Teaching Assistant @ Indian Institute of Technology Madras
Jan 2020 → May 2020
Indian Institute of Technology Madras
- Designed and evaluated theoretical and practical assignments on various topics in Natural Language Processing.
- Presented lectures on Edit Distance and the Cocke-Young-Kasami (CYK) algorithm, to a class of 70 students.
- Mentored sixteen pairs of students on research projects, with supervision through regular team-wise progress meetings.
Projects
Graph Neural Networks for Extreme Summarization
Indian Institute of Technology Madras
- Formulated appropriate graph-based deep neural models for the Extreme Summarization (XSum) task with sentence-level and/or document-level graphs, and obtained better performance than simple recurrent and hierarchical models.
Risk-Sensitivity in Multi-Armed Bandits
Indian Institute of Technology Madras
- Surveyed and implemented risk-sensitivity methods for stochastic bandit problems, and upgraded the Explore-Then-Commit algorithm for VaR and cVaR measures with competent performance.
Leveraging Ontological Knowledge for Neural Language Models
Indian Institute of Technology Madras
- Incorporated Weight Initialization in learning word embeddings using the WordNet Ontology for a task in the Construction domain, resulting in a faster convergence rate and better representation of domain-specific terms.
Multimodal Dialogue Generation
Indian Institute of Technology Madras
- Developed a deep neural model to establish the positive effect of domain features in the performance of image retrieval in multimodal dialogue systems and explored the performance of attention and memory-based models with adaptations for multimodal dialogue and domain knowledge integration.
Risk-Sensitive Reinforcement Learning
Indian Institute of Technology Madras
- Empirically analyzed the existing methods for risk-sensitive reinforcement learning, tested the effectiveness of modified versions and proposed a new distance-based risk measure and algorithm for Gridworld.
Summarization and Keyword Extraction using TextRank
Indian Institute of Technology Madras
- Analysed the TextRank algorithm for keyword extraction with syntactic filters and augmentation via Explicit Semantic Analysis, and for text summarization with exploration of various textual similarity methods.
Scaling Graph Algorithms
Indian Institute of Technology Madras
- Implemented optimized graph algorithms for maximum network flow and finding a maximum matching in a bipartite graph for real data graphs with up to 10,000 vertices and 100,000 edges.
Skin Disease Diagnostic System
Microsoft code.fun.do Contest, Indian Institute of Technology Madras
- Designed a web application that attempts to diagnose skin diseases based on images of the user's skin powered by a deep neural model trained on a dataset created by scraping images from the web.
Breakout Game
Indian Institute of Technology Madras
- Developed an Android application for the Breakout game with basic playing and scoring features.
Skills
CC++C#JavaPythonHTMLCSSJavascriptWeb AssemblyUnityARCoreAndroid StudioStanford CoreNLPGitBootstrapjQueryEmscriptenBlazej2CINLTKdjangoscipypandassklearngensimkerastensorflowpytorch
Education
Indian Institute of Technology Madras, Chennai, India
2015 → 2020
Dual Degree (B.Tech + M.Tech), Computer Science and Engineering
KLE Society's Independent PU College, Bangalore
2015 → 2015
XII - Karnataka Board
97.30%
B P Indian Public School, Bangalore
2013 → 2013
X - ICSE
96.33%