Ananjan Nandi

sequoia_pic.jpg

Second Year MSCS Candidate

Stanford University

Welcome to my site!

I am a second year MS Computer Science candidate at Stanford University. I was previously an undergraduate student at the Department of Computer Science, IIT Delhi. My current research interests include improving the generalization and reasoning capabilities of Large Language Models, with a focus on interpretable approaches such as algorithmic distillation. I have also worked on LLM safety, chatbot design, human-computer interaction, knowledge graph completion, remote sensing, drug discovery and extreme classification.

I am a Research Assistant at the Stanford NLP Group under the supervision of Prof. Chris Manning, building a method to induce syntactic inductive biases in transformer language models without any architectural modifications. I have also worked with Prof. Dan Jurafsky, Prof. Karen Livescu, Prof. Diyi Yang and Prof. Tatsunori Hashimoto on various NLP-related projects over the past year. These include a domain-specific language for jailbreak attack synthesis, a robust optimization method to boost the performance of worst-performing langauges while training multilingual speech recognition models, and a chatbot that enables domain experts to tailor LLM-based psychotherapy patient simulations for roleplay practice.

I have previously worked on augmentation and ensembling techniques for Knowledge Graph Completion at the Data Analytics and Intelligence Research group under the supervision of Prof. Mausam and Prof. Parag Singla. I have also worked on the generation of Land Use-Land Cover maps from satellite time series data in the Appropriate Computing Technologies for Development group under the supervision of Prof. Aaditeshwar Seth.

I interned at Palantir Technologies Seattle as a ML Researcher Intern over the summer. I worked on semantic parsing of natural language queries into executable statements in an internal query language, as well as an enterprise CoPilot. I also did a 7-month stint as a part-time AI researcher intern at KnowDis Data Science, where I worked on problems related to multilingual idiom alignment, spelling-robust language identification, extreme classification, and drug discovery. I also interned at D.E. Shaw India as a Member Technical Intern in the Strategy Tech department at the end of my third year in college. I have played table tennis/ping pong for more than a decade and cricket at the Under-14 and Under-17 levels, and also enjoy quizzing and video games.

news

Nov 28, 2024 Sneaking Syntax into Transformer Language Models with Tree Regularization” now available on ArXiv!
Sep 20, 2024 My paper “Roleplay-doh: Enabling Domain-Experts to Create LLM-simulated Patients via Eliciting and Adhering to Principles” was accepted as an oral presentation at EMNLP 2024!
Sep 13, 2024 h4rm3l: A Language for Composable Jailbreak Attack Synthesis” now available on ArXiv!
Jun 17, 2024 Started my ML Researcher Internship at Palantir Seattle!
May 16, 2024 My paper “DynaSemble: Dynamic Ensembling of Textual and Structure-Based Models for Knowledge Graph Completion” was accepted as a poster at ACL 2024!