Sid Srikanth

I'm an undergraduate researcher at the University of Southern California in Los Angeles. I am grateful to be advised by Stefanos Nikolaidis at the ICAROS Lab, and John Krumm and Cyrus Shahabi at the InfoLab. I have previously worked with Ashish Amresh at the Decision Theater.

I am actively applying to PhD programs starting in Fall 2026!

Scholar  /  X  /  GitHub  /  LinkedIn

Research

My high-level research interests are using robot planning and learning to build generalizable robots and improving human-robot interactions. Throughout my undergrad, I have explored facets of this problem through the lens of quality diversity optimization and scenario generation. I also conduct research in causality inference and path planning using geospatial data.

Past work

We introduce Q-DIG, a novel red-teaming framework for generating diverse task instructions that induce failures in SoTA Vision-Language-Action models using Quality Diversity. We then provide a methodology to leverage these failure modes to improve policies.

We introduce PLAN-QD: a framework that algorithmically generates a population of diverse, human-like teaming and communicative behaviors by leveraging LLM-powered agents and Quality Diversity, an evolutionary learning framework.

We present NEXICA, an algorithm to discover which parts of a highway system tend to cause slowdowns on other parts of a highway. We evaluate our approach on 6 months of timeseries data on 195 stations in Los Angeles, and show that it can efficiently discover causality relationships that are consistent with human intuition.