I am a Master's student in Electrical and Computer Engineering at the University of Michigan, Ann Arbor. I build ML models, physics-informed computational models/simulations, and robotics and optimization algorithms for real-world systems — from battery characterization and aging prediction to electromagnetic scattering simulations, autonomous robot navigation, and 3D scene understanding.
I work with Prof. Ziyou Song at the Electric Vehicle Center on physics-informed EIS models and ML-based battery aging prediction. I work with Prof. Raj Rao Nadakuditi on differentiable physical simulations and gradient-based optimization for electromagnetic scattering systems. I also work at the HDR Lab on real-time 3D depth estimation for visuo-tactile sensors on robotic arms.
Before joining UMich, I completed my B.Tech in Electrical Engineering from IIT (ISM) Dhanbad, building a foundation in machine learning, computer vision, and robotics software. I am actively looking for summer internship opportunities.
Building deep learning models for real-world prediction — attention-enhanced Seq2Seq models for time-series forecasting, multimodal model fine-tuning, and end-to-end ML pipelines on large-scale datasets
Coupling first-principles equations with data-driven methods — physics-informed EIS models for battery characterization, differentiable physical simulations with gradient-based optimization for inverse design
Applying large foundation models to physical robot systems — fine-tuning vision-language-action models (OpenVLA) for robot manipulation, and integrating vision foundation models (Grounding DINO) with SLAM for language-guided autonomous navigation
Applying deep learning for object detection, instance segmentation, depth estimation, and multi-view 3D scene reconstruction in scientific and engineering applications
Working on hybrid physics-based EIS model paper draft for SEI characterization in Li-ion batteries (with Prof. Ziyou Song)
Started working with Prof. Raj Rao Nadakuditi on translating and understanding electromagnetic scattering simulation codebase
Joined the Electric Vehicle Center (Prof. Ziyou Song) working on physics-based battery modeling and ML-based aging prediction
Started MS in ECE at University of Michigan
Electric Vehicle Center, University of Michigan — Prof. Ziyou Song
Built physics-informed EIS models for Li-ion battery characterization, extracting degradation parameters (SEI, LAM, LLI) from BOL to EOL via Differential Evolution optimization — achieving 1.7% mean error across 117 spectra. Developed attention-enhanced Seq2Seq models for battery aging prediction, achieving 2–5% MAPE on capacity-fade trajectories from the first 30% of cycling data across 225 cells. Designed accelerated test protocol frameworks converting real-world driving profiles to battery power demand cycles.
University of Michigan — Prof. Raj Rao Nadakuditi
Translating the CyScat electromagnetic scattering simulation codebase from MATLAB to Julia and Python; implemented differentiable physical simulation via forward-mode automatic differentiation through the T-matrix solver, enabling gradient-based optimization to find optimal refractive index and wavelength configurations for maximum wave transmission through photonic structures. Validated numerical precision against finite-difference checks to machine precision.
HDR Lab, University of Michigan
Developing real-time 3D depth estimation algorithms for visuo-tactile sensors on robotic arms. Built an iOS app for automated hardware data collection enabling large-scale dataset generation for model training and evaluation.
Carnegie Mellon University — Dr. Arun Balajee Vasudevan
Engineered end-to-end data generation and labeling pipelines to curate a 1M+ sample dataset of real/synthetic audio pairs. Fine-tuned CLAP audio-language model and Llama-8B for deepfake speech detection, achieving 90% accuracy. Paper submitted to Interspeech 2026 (under review).
Mowito
Developed Multi-head Mask R-CNN models for instance segmentation and object detection in warehouse environments; integrated into production robot perception pipelines. Implemented perceptual hashing for large-scale dataset deduplication (30% size reduction) and built order-fulfillment optimization using Leiden clustering.
Hybrid physics-based SEI impedance model with degradation-constrained differential evolution optimization, achieving 1.7% mean error across 6 cells and 117 EIS spectra
Learn More →Translation of EM scattering simulation codebase from MATLAB to Julia/Python, with wavefront optimization via scattering matrix eigen-channel analysis
Learn More →Attention-enhanced sequence-to-sequence models for battery capacity-fade trajectory prediction from early cycling data, with 2–5% MAPE
Learn More →Matrix Methods for Machine Learning and Signal Processing
Probability and Random Processes
