Greater Chicago Area
I'm David, an Machine Learning Researcher at Jane Street. Before that, I was a PhD Student in machine learning at TTIC, happily advised by Matthew Walter. I worked with other senior members of TTIC and UChicago CS including Michael Maire and David McAllester. During my PhD I had the pleasure of interning at NVIDIA (remote) with Rui Shen; at IBM Research (New York) with Danish Contractor, Chulaka Gunasekara, and Pavan Kapanipathi; and at Google DeepMind (London) with Fantine Huot, Michael Kaisers, and Mirella Lapata. I was supported by a 2019 NSF Graduate Research Fellowship. I have broad experience across empirical deep learning, with focus in language modeling, reinforcement learning and the optimization behavior of neural networks. See my website (dyunis.github.io) for more details. Before my PhD, I graduated from the University of Chicago with a BS in Mathematics and a BS in Molecular Engineering. During that time, I spent a few years working with Matthew Walter at TTIC.
researchey things
Worked on multiturn reinforcement learning with language models
- Designed and led a project on scoping billion-parameter language models for specific uses - Scaled thousands of runs across multiple datasets for detailed ablation on existing alignment techniques