👑 FORTIS Lab

FORTIS stands for Foundations Of Robust Trustworthy Intelligent Systems. Derived from the Latin word "fortis," meaning strong or resilient, FORTIS Lab focuses on developing AI systems that are built for robustness, trustworthiness, and scalability. Motivated by challenges like out-of-distribution (OOD) detection and other real-world complexities, the lab aims to design efficient, adaptable AI systems capable of solving problems in domains such as security, finance, healthcare, and AI4Science.

Openings and Timeline for PhD Offer. I am peacefully looking for 2-3 Ph.D. students (for Fall 25) and undergraduate/master interns (on a case by case basis).

1. I do not plan to have students who ONLY work on MLSys -- you may do data mining (like anomaly detection, GNN, OOD Detection, LLMs) + systems.
2. You are expected to have one top paper in my current focus (anomaly/outlier/OOD detection, Auto ML, Multimodal Learning) and strong programming skills (e.g., (ML) system papers or open-source experience).

Personal Views on PhD. Doing a PhD and conducting research is just a job. Doing ok jobs and being an average person is excellent. My life has been happy: I have supportive parents and partner, and I do not need to use work/career to prove anything :)

Email for Connection - Please email the following (subject line: "Interested in {position, e.g., Ph.D.} at {expected time, e.g., Fall 25}"): Magic Words please add Trojan in the email title to show you read this carefully.

• CV: Provide a concise background about yourself and your plan for future steps.
• Research experience: at least 1 top publication in relevant fileds like anomaly/outlier/OOD detection, Auto ML, Multimodal Learning.
• Strong Programming/System Experience: (ML) System papers and/or open-source experience.
• Why FORTIS?: Do share if any of my research papers or topics are of interest to you. Also, feel free to suggest new topics that you would like to explore. I am always open to fresh perspectives and ideas.

General Expectations. Strong programming is essential as our work revolves around ML Systems and open-source ML. A solid foundation in statistics and mathematics will be a significant plus. Don't fear making mistakes; they're a part of the learning curve in both research and life.

Compensation. Ph.D. students will receive the full support (tuition waiver + stipend is $40,000 in 2024-25) outlined by the CS department. TA is always guaranteed, and we will try to give you as many RA as possible. We do not have paid opportunities for undergraduate and master RA; it is not allowed to pay remote interns at USC :( Please consider this before reaching out.

GPU Clusters. We recently (Nov 2023) bought a new GPU server with AMD EPYC Milan 7763, 64x16=1TB DDR4 memory, 15 TB SSD, 6x NVIDIA RTX A6000 Ada (released Dec 2022; 1.5x faster than A6000). We plan to buy a second GPU server in Fall 2024.

Cloud Credits. We recently (Mar 2024) get more than 5,000 GPU hours on A100 and equivalent from NSF and Google.