Computer Scientist III · NASA ODSI

Pranath Reddy
Kumbam

I build AI systems for scientific work.

At NASA, I build multimodal foundation models and agentic systems for scientific research. I also lead DeepLense, where we develop physics-aware machine learning for gravitational lensing.

Pranath Reddy Kumbam outdoors
Huntsville, Alabama · 2026

I focus on AI systems that scientists can inspect, evaluate, and use in real research workflows.

01

Scientific agents

Agents that search evidence, plan experiments, and preserve the context needed for long research tasks.

02

Foundation models & open science

Multimodal models, grounded research knowledge systems, and tools that make scientific code easier to discover and reuse.

03

Physics-aware learning

Models that use symmetry, geometry, equations, and simulation structure to learn from physical data.

Papers

Peer-reviewed papers, workshop papers, and preprints across agents, AI for science, and astrophysics. My name is highlighted.

Google Scholar ↗
2026arXiv · NASA

Scientific Code Search at Scale: A Multi-Domain Dataset and Benchmark

Nishan Pantha, Pranath Reddy Kumbam, Sajil Awale, Pushwitha Krishnappa, Muthukumaran Ramasubramanian, Nidhi Jha, Emily Foshee, Ankur Kumar, Rachel Slank, Ashkbiz Danehkar & Rahul Ramachandran

Read ↗
2026ICML

Selective Memory Retention for Long-Horizon LLM Agents

Pranath Reddy

Read ↗
2026AAAI FSS · Spotlight

DLens: Closed-Loop AI Agents for Parsimonious Scientific Machine Learning in Dark Matter Gravitational Lensing

Aatmaj Amol Salunke, Mywish Anand, Pranath Reddy Kumbam, Michael W. Toomey & Sergei V. Gleyzer

Paper ↗
2026NeurIPS DiffuLM

Trace–Answer Compatibility Emerges at Depth and Mediates Prediction in Diffusion Language Models

Ashutosh Ojha, Pranath Reddy & Sergei Gleyzer

PDF ↗
2026NeurIPS RPS

Lens-LeJEPA: Learning the Right Invariances for Strong Gravitational Lenses

Arnesh Batra, Karthik Gaur, Pranath Reddy, Michael W. Toomey & Sergei Gleyzer

PDF ↗
2026ML4PS

A Polar Coordinate Prior for Self-Supervised Strong-Lens Representations

Karthik Gaur, Pranath Reddy, Michael W. Toomey & Sergei Gleyzer

PDF ↗
2026ML4PS

VELA: Stationary Latent Targets for Label-Efficient Strong-Lensing Representations

Karthik Gaur, Pranath Reddy, Michael W. Toomey & Sergei Gleyzer

PDF ↗
2026ML4PS

Bilinear C4-Covariant Spectral Neural Operator (BiC4-NO) for Scientific Field Surrogates and Gravitational Lensing

Sushmanth Reddy, Pranath Reddy, Michael W. Toomey & Sergei Gleyzer

PDF ↗
2026ML4PS

Pooling Discards Resolution-Invariant Structure: An Architecture and Ablation Study on Darcy Flow

Paras Balani, Pranath Reddy, Michael W. Toomey & Sergei Gleyzer

PDF ↗
2026AAAI FSS

D4-Equivariant Hybrid Quantum–Classical Learning for Strong Gravitational Lens Classification

Sushmanth Reddy, Pranath Reddy, Michael W. Toomey & Sergei Gleyzer

Paper ↗
2025ApJ

Domain Adaptation in Application to Gravitational Lens Finding

Hanna Parul, Sergei Gleyzer, Pranath Reddy & Michael W. Toomey

DOI ↗
2025ICWSM

Exploiting Explainability to Design Adversarial Attacks and Evaluate Attack Resilience in Hate-Speech Detection Models

Pranath Reddy Kumbam, Sohaib Uddin Syed, Prashanth Thamminedi, Suhas Harish, Ian Perera & Bonnie J. Dorr

DOI ↗
2025NeurIPS ML4PS

FlowLensing: Simulating Gravitational Lensing with Flow Matching

Hamees Sayed, Pranath Reddy, Michael W. Toomey & Sergei Gleyzer

Read ↗
2025NeurIPS ML4PS

Lens-JEPA: Physics-Informed Joint Embedding Predictive Architecture for Gravitational Lensing

J. Rishi, Pranath Reddy, Michael W. Toomey & Sergei Gleyzer

PDF ↗
2025NeurIPS ML4PS

HEAL-PINN: Physics-Informed Swin Transformer for Sparse Lensing Data

Dhruv Srivastava, Pranath Reddy, Michael W. Toomey & Sergei Gleyzer

PDF ↗
2025NeurIPS ML4PS

From Simulation to Survey: Benchmarking Super-Resolution for LSST-like Lensing Data

Aleksandr Duplinskii, Pranath Reddy, Michael W. Toomey & Sergei Gleyzer

PDF ↗
2024MLST

DiffLense: A Conditional Diffusion Model for Super-Resolution of Gravitational Lensing Data

Pranath Reddy, Michael W. Toomey, Hanna Parul & Sergei Gleyzer

DOI ↗
2024NeurIPS ML4PS

LensPINN: Physics-Informed Neural Network for Learning Dark Matter Morphology in Lensing

Ashutosh Ojha, Sergei Gleyzer, Michael W. Toomey & Pranath Reddy

PDF ↗
2024arXiv

FloodLense: A Framework for ChatGPT-based Real-time Flood Detection

Pranath Reddy Kumbam & Kshitij Maruti Vejre

Read ↗
2023ApJ

Domain Adaptation for Simulation-Based Dark Matter Searches Using Strong Gravitational Lensing

Stephon Alexander, Sergei Gleyzer, Pranath Reddy, Marcos Tidball & Michael W. Toomey

DOI ↗
2021MLST

A Hybrid Quantum Regression Model for the Prediction of Molecular Atomization Energies

Pranath Reddy & Aranya B. Bhattacherjee

DOI ↗
2020NeurIPS ML4PS

Decoding Dark Matter Substructure without Supervision

Stephon Alexander, Sergei Gleyzer, Hanna Parul, Pranath Reddy, Michael W. Toomey, Emanuele Usai & Ryker Von Klar

Read ↗
2020MLST

Baryon Density Extraction and Isotropy Analysis of Cosmic Microwave Background Using Deep Learning

Amit Mishra, Pranath Reddy & Rahul Nigam

DOI ↗

I work across research and software engineering.

I am a Computer Scientist III at NASA’s Office of Data Science and Informatics, where I work on foundation models, scientific code discovery, and agentic research systems. Before NASA, I worked on gravitational-lensing research at the University of Alabama, NPU compiler software at Samsung Research, and open-source machine learning through Google Summer of Code and CERN-HSF.

Current
Computer Scientist III · NASA ODSI
Education
M.S. Computer Science, University of Florida
M.S. Physics + B.E. EEE, BITS Pilani
Recognition
AAAI Spotlight · ICML Gold Reviewer · Apple Swift Student Challenge Winner
Based in
Huntsville, Alabama

Have a research question or collaboration in mind?