I am a 2nd year PhD student in Computer Science at Georgetown University, and working on Natural Language Processing (NLP) advised by Prof. Nathan Schneider.
My research focuses on challenges of computation, data, and linguistics in language processing with an eye towards building better systems and technologies; ultimately in service of the languages of the world and the communities that use them.
At Georgetown, I am an active member of
- Nathan’s Excellent Research Team (NERT)
- Corpling led by Amir Zeldes
- GUCL (Computation and Language @ Georgetown)
My recent projects span:
This work has been thankfully informed by Legal Scholarship, Cognitive Science, Linguistics, and Humanities at Georgetown and beyond.
I am currently working on (even more) projects on metalanguage, machine translation, and discourse.
News
- 2026-02 : Prof. Nathan Schneider presented our work on LLM Legal Interpretation at SOLID 2026 at Stanford Law School.
- 2025-10 : Two workshop items at EMNLP 2025.
- On Legal interpretation with LLMs in NLLPW 2025, presented by Junghyun Min. (abstract preprint)
- DeDisCo at the DISRPT 2025 Shared Task in CODI-CRAC 2025, presented virtually by Zhuoxuan Ju. (paper demo)
- 2025-09 : DeDisCo, led by Zhuoxuan Ju and Jingni Wu, and mentor Amir Zeldes won the 2025 DISRPT Shared Task on Discourse Relation Classification across Formalisms. (demo, news)
- 2025-05 : Fortunate to close my first academic year at Georgetown CS with the Outstanding TA Award.
- 2025-03 : It was an honor to be nominated for Graduate Student Teaching Assistant Award in Georgetown University’s 2025 Graduate Student Awards.
Publications
Submissions (Under Review, not preprints yet)
- Abhishek Purushothama*, Emma Thronson*, Alexia Guo, Amir Zeldes. 2026. Universal Dependencies for In-Context Coptic Translation
- Dawson Petersen, Abhishek Purushothama, Nathan Schneider. 2026. Sense and Sensitivity: “Reasoning” Models are More Robust, but can Diverge from Human Consensus in a Legal Interpretation Task
Coverage
-
‘DeDisCo’, our winning system in 2025 DISRPT Shared Task on Discourse Relation Classification across Formalisms
By Georgetown Grad Studies, By ‘The Hoya’, By Georgetown Linguistics
Languages
Luke Gessler said that “linguists are often asked how many languages they speak” and now that I work with them a lot, here are mine:
- Kannada (kan)(native, ~B1 level)
- Hindi (hin)(intermediate, ~B1 level)
- English (eng)(professional, ~C1 level),
- Spanish (spa)(a single university course, ~A1 level).
My current language goals are to be able to
- converse and read (a specific novel) in Spanish
- read Konkani (kok) with Kannada script.
- read and work with Coptic (cop).
Similarly, if you ask me how many programming languages I can program in, I will say (like I tell my students)
“it’s not about the number of programming language(s), it’s about the principles”
but then go on to say I have programmed in many languages:
Python (research and tooling), Java (enterprise cloud software), Scala (TAing and side projects, linkedin assessment), C (let’s say structs and pointers), Javascript (up to JQuery, +Angular, Node), Scheme, OCaml (coursework) and Kotlin (graduate software engineering project).
Next Moving Targets: R has priority now over Julia for visualization. More proficiency in OCaml because of Grew.
Writing
I write technical and non-technical blog articles infrequently, you can find them on my medium site here.
Bookshelf
I like to read (largely fiction) and for better or worse track it on Goodreads. If you are interested you can find my shelves here. If you ever meet me, feel free to ask for a personalized recommendation :).
Some lists
I maintain a few (interesting) non-technical lists for fun! I made them separate sites to make space this page compact.
Source
Xiulin Yang’s spiffy site made me move from my old site (archiving in progress…) to this new site built with Rayeren’s acad-homepage. My code is available in Github.