Meet the Patronus team: Jitin Krishnan

Bio:

Jitin was a Research Scientist at Meta, where he specialized in post-training and alignment of large language models (LLMs) for conversational commerce applications. In 2024, he led the end-to-end design and deployment of fine-tuned multilingual LLaMA models, enhancing response quality across businesses in Thai, Spanish, and Portuguese. From 2022 to 2023, as part of the Modern Recommender Systems organization, he developed cross-domain user-understanding embeddings. These embeddings improved ranking and retrieval performance across Facebook and Instagram's recommendation systems, notably increasing session duration, watch time, and daily active users, especially among cold-start cohorts. Earlier, he contributed to building integrity embeddings to support Meta's integrity teams in identifying sensitive, dangerous, and other harmful content. Prior to this, he interned at Amazon Alexa and NASA Goddard, where he contributed to enhancing natural language understanding (NLU) capabilities for virtual assistants. He earned his Ph.D. in Computer Science from George Mason University in 2021, focusing on cross-lingual and cross-domain transfer learning. He holds a B.S. in Computer Science with a Minor in Economics from the University of Virginia.

Q: Why are you excited to join Patronus AI?‍

I'm most excited about doing cutting-edge applied research on LLMs focusing on agentic AI, alignment, evaluation, and scalable oversight. I'm particularly interested in developing new methods for aligning agentic models with human intent in open-ended environments, and building evaluation frameworks that can robustly measure model behavior under distribution shift. I’m also passionate about making scalable oversight techniques more practical — especially using synthetic data, automated red-teaming, and reward model distillation. Beyond the technical challenges, I'm eager to collaborate with a talented and passionate team of engineers and researchers who share a drive to push the boundaries of what safe and capable AI systems can do.

‍Q: What do you like to do in your free time?

Outside of work, I love to travel and explore new cultures. I try to visit at least one new country every year. Each year, I also embark on long motorcycle journeys, exploring scenic routes like California’s Route 1, circling the Olympic Peninsula, and riding through Glacier National Park. To stay active, I enjoy playing soccer and badminton.