Soham Palande
Research at J.P. Morgan AI Research

Hi, I’m Soham. I’m currently a research engineer in J.P. Morgan’s AI Research group.
Research: I study how machine and human models of cognition exhibit strategic reasoning in structured, multi-step environments—and their emergence through behavioral evaluation, learned representational structures (from language and fine-tuning), and mechanistic insight in stylized decision-making scenarios involving uncertainty, adaptation, and long-term planning—such as bidding, bluffing, or probabilistic inference—where optimal strategies must be inductively discovered rather than precomputed.
My goal is to understand when and how models demonstrate intelligent, goal-directed behaviors, and precisely which internal structures—learned, emergent, or structurally implicit—support these capabilities. This involves behavioral analysis, algorithmic investigation, and mechanistic insight, drawing from game theory, cognitive modeling, and computational geometry. I’m particularly interested in how adaptive reasoning naturally develops in response to incentives, reinforced feedback signals, inductive contexts, and the geometric structure of internal language representations.
Ultimately, through empirical evaluations, my work addresses deeper theoretical questions—such as how models inductively simulate, generalize, or abstract analogously to human cognition in scientific and mathematical reasoning. By studying reasoning within computational systems and natural cognition through a computational lens, I aim to uncover generalizable principles of intelligent behavior, enabling the design of more powerful and adaptive systems.
Previously: I graduated from Rutgers University-NB majoring in Computer Science and Mathematics. I was an R&D intern at L3Harris where I worked on building anomaly detection models.
news
*not updated
May 10, 2022 | I was awarded the Chancellor-Provost’s Research Excellence Award |
---|---|
May 5, 2022 | I was awarded the Nicholas Novielli Prize for academic excellence by the Rutgers CS Dept. |