Joseph Selvaraaj

My broad research interests focus on advancing multi agent autonomous systems by designing algorithms and systems that are

1. theoretically grounded and efficient under different notions of efficiency,
2. scalable to complex problems like sequential social dilemmas,
3. simple to understand and implement.

To that end, I am exploring how deep learning, deep RL, deep MARL as a tool can be combined with insights from pure and applied mathematics.

---

Bio: I am currently a software engineer in Lucid Software’s Cards 2 team. Alongside my work, I also pursue independent research on multi agent reinforcement learning. I got my B.S computer science, B.S mathematics and B.B.A finance from UMass Amherst in 2023.

At UMass, I studied applying neural differential equations for episode trajectory modeling of grid environment as part of my undergraduate honors thesis with Prof. Bruno Castro da Silva.

---