Incoming MSCS · NYU Courant · Fall 2026
From ML at scaleto embodied AI.
I’m an incoming MSCS student at NYU Courant (Fall 2026). I come from several years of building machine learning systems at scale, and I’m now focused on embodied AI — bringing modern ML/DL to agents that have to act in the physical world.
What I’m working toward. I want to study how the methods that made large-scale ML work — reinforcement learning, large pretrained models, and learning from messy real-world data — transfer to robots. I’m especially interested in:
- Reinforcement learning for robot control and manipulation
- Test-time scaling for embodied agents — spending inference-time compute and search to make robot policies more capable and reliable
- The intersection of large pretrained / foundation models (vision-language-action models, world models) and embodied decision-making
Where I’m coming from. Most recently I was an Applied Scientist at Amazon, training transformer-based small language models for multilingual search query understanding in production. Before that, as a Research Engineer at IBM Research, I worked on end-to-end deep reinforcement learning for conversation disentanglement (AAAI 2023), domain-specialized language models for logs, and large-scale log analytics — work published at ICSE, AAAI, IEEE CLOUD, and ICPE, with several granted U.S. patents. I did my B.Tech in Computer Science at IIT Hyderabad (2018–2022), including research on autonomous-driving benchmarks in Lab 1055.
The common thread across everything I’ve done is turning research into something that actually runs. I want to do that for embodied systems next.
You can reach me at choudhary121karan@gmail.com, read my CV, or browse publications and patents.
News
- Fall 2026Starting the MSCS program at NYU Courant; moving toward embodied AI.
- Nov 2025Paper on scalable LLM-based log analysis accepted at AAAI 2026.
- Sep 2025Patent US12536061B1 granted (explainability-driven alert suppression).
- Apr 2025Patent US12277025B2 granted (dynamic alert-noise reduction).
- Oct 2024Joined Amazon as an Applied Scientist on multilingual search query understanding.