CV

PhD student

Technische Universität Darmstadt

PhD student at Technische Universität Darmstadt, advised by Angela Yu in the Computational Modelling of Intelligent Systems lab.

Research Trainee

Basis, New York City, NY

Contributing to Project MARA (Modeling, Abstraction, and Reasoning Agents).

Research Assistant

DFKI, Kaiserslautern, Germany

Conducted research on probabilistic inference methods and Bayesian machine learning. Served as teaching assistant for the ``Machine Learning in Julia’’ course.

Short-term Research Scholar

Columbia University, New York City, NY

Developed a framework that automatically constructs simplified model abstractions to balance computational accuracy with resource constraints.

Short-term Research Scholar

Columbia University, New York City, NY

Implemented parametric inversion in JAX — a program transformation technique that inverts non-injective functions by introducing parameters to uniquely determine inputs. Demonstrated approach on simulation models, including Lotka-Volterra dynamics.

Student Developer

Google Summer of Code

Developed CounterfactualFairness.jl package for counterfactual reasoning methods for algorithmic fairness. Implemented multiple fairness criteria, including counterfactual fairness and path-specific effects.

Content and Publicity Head

IEEE Student Branch, BIT Mesra, India

Conducted workshops and organized technical events for the student community.

Mentor

MTTS Program

Supported students in the Level 0 advanced mathematics program.

Master of Science

Technische Universität Kaiserslautern (RPTU)

Computer Science

• Machine Learning I & II
• Monte Carlo Algorithms
• Automated Reasoning
• Functional Programming
• 3D Computer Vision
• Stochastic Modeling of Cognitive Processes

Foundational Course in Data Science

Indian Institute of Technology Madras

Bachelor of Science

Birla Institute of Technology, Mesra

Mathematics and Computing

• Linear Algebra
• Discrete Mathematics
• Real and Complex Analysis

Top 300, National Programming and Data Science Qualifier

IIT Madras

Achieved 95% score in the national programming and data science qualifier.

Letter of Appreciation

Union Minister of India, Smriti Irani

Received a letter of appreciation from the Union Minister of India.

Bronze Medal, International Olympiad of Mathematics

International Mathematics Olympiad

Awarded bronze medal in the International Olympiad of Mathematics.

Benchmarking World-Model Learning with Environment-Level Queries

Forty-third International Conference on Machine Learning

We introduce WorldTest, a framework for evaluating world model learning through environment-level queries about full dynamics, not single trajectories. AutumnBench, instantiated under WorldTest, exposes large human-LLM gaps in world-model building.

Had Enough of Experts? Elicitation and Evaluation of Bayesian Priors from LLMs

NeurIPS 2024 Workshop

Co-authored work on eliciting and evaluating Bayesian priors from Large Language Models.

AutumnBench

Developed a framework and becnhmark for evaluating world model building in interactive agents.

• World model evaluation
• Interactive agents
• Benchmarking framework

Rational Abstraction

Developed a framework for rational abstraction, a technique for creating simplified model abstractions to balance computational accuracy with resource constraints.

• Model abstraction
• Resource-rationality

CounterfactualFairness.jl

Developed a Julia package for counterfactual reasoning methods for algorithmic fairness, implementing multiple fairness criteria.

• Algorithmic fairness
• Counterfactual reasoning
• Julia package

ProbMods in Omega.jl

Developed a translation of ProbMods (probmods.org) in Omega.jl, a probabilistic programming language.

• ProbMods
• Probabilistic programming
• Omega.jl

Parametric Inversion in JAX

Implemented a program transformation technique that inverts non-injective functions by introducing parameters to uniquely determine inputs.

• Program transformation