I am a PhD from the Johns Hopkins University's CLSP (Center for Language and Speech Processing), where I worked on machine learning and natural language with interests in typologically diverse languages, various kinds of structured prediction, establishing and using clean formalisms, and appreciating neat mathematical models of interesting phenomena and solid engineering.

These days I work with the most amazing people at Genentech (subsidiary of Roche) in its Prescient Design accelerator right in Midtown Manhattan, helping leverage state-of-the-art machine learning (and LLM technology in particular) to enable groundbreaking drug discovery work!

Recent blog entries

• From PyTorch to JAX: towards neural net frameworks that purify stateful code (2020-03-09)
• Clowning in Pennsylvania (2020-03-02)
• My name is Sabrina – I'm trans (2019-12-30)
• Can you compare perplexity across different segmentations? (2019-04-23)
• NLP/CL Twitter Megathread (2017-04-08)
• Language diversity in ACL 2004 - 2016 (2016-12-22)
• Describing discontinuous constituents with LCFRS (2016-10-21)

CV (the gist of it)

Machine Learning Engineer, Large Language Models, Genentech, New York, NY, USA (06/2024 - current)

• work as part of the Prescient Design Accelerator/MLDD (machine learning drug discovery) unit in gCS (Genentech Computational Sciences), part of gRED (Genentech Research and Early Development) in Genentech, which is a subsidiary of Roche’s Pharma division
• collaborating with biomedical teams in deciding how to bring large language models into applications from foundations to clinical outcomes

Senior AI Research Engineer, AlphaSense, New York, NY, USA (11/2023 - 06/2024)

• work on AlphaSense Assistant, an AI-powered chatbot designed to answer business questions
• primarily working around the task of “citation,” relating chatbot generations to underlying data in a model-agnostic way to aid verifiability of results by platform and user
• extensive visualization efforts for citations and metadata-based, neural, and non-neural statistical signals, communicating results to leadership to drive key decisions in the rollout of the AI-powered Assistant model

Ph.D. Computer Science, Johns Hopkins University, Baltimore, MD, USA (09/2017 - 10/2023)

• advised by Prof. Jason Eisner at the CLSP (Center for Language and Speech Processing)
• Thesis: “Building and evaluating open-vocabulary language models”

Internship, Cohere AI, Toronto, Canada (summer of 2022)

• engineering a framework for the scalable training of large language models using JAX and TPUs

Part-time internship, HuggingFace, New York City, NY (summer of 2021)

• tokenization in language modeling

Internship, Facebook AI Research (FAIR), New York City, NY (summer of 2020)

• analyzing (chit-)chat bots, using metacognition to improve linguistic calibration

Internship, Google, New York City, NY (summer of 2019)

• transliteration using pronunciation data and cross-language multi-task approaches

M.Sc. Computer Science, TU Dresden, Germany (finished 07/2017)

• thesis: “Soft matching of terminals for syntactic parsing” (supervised by Prof. Heiko Vogler)

Internship, USC Information Sciences Institute, Los Angeles, CA (summer of 2016)

• machine translation / DARPA LORELEI under Prof. Kevin Knight and Prof. Daniel Marcu
• translating out-of-vocabulary words with unprocessed human-readable dictionaries

B.Sc. Computer Science, TU Dresden, Germany (finished 08/2015)

• thesis: “Extracting and Binarizing probabilistic linear context-free rewriting systems”