Academic Experience
Purdue University · AIDA3 Research Center
Graduate Research Fellow
West Lafayette, Indiana
- Research anomaly detection for engine health monitoring using variational autoencoders for sensor reconstruction and interventional simulation experiments.
- Developed deep-learning computer vision algorithms to estimate human mental states using the Theta-Alpha ratio (TAR), an EEG-derived ground truth, and exogenous ocular features.
- Created a novel wavelet-based transformer for accurate autoregressive prediction of TAR.
- Published NARCOX: Non-Stationary Auto-Regressive Transformers for Cognitive Modeling with Ocular Exogenous Input in Neurocomputing.
Skills: Deep Learning · PyTorch · Time Series Modelling · Cognitive Load Theory
UCLA · Laboratory for Robust Information Systems (LORIS)
Graduate Student Researcher
Los Angeles, California
- Researched quantum key distribution and secure information reconciliation between two users.
- Used LDPC codes, Slepian-Wolf coding, and a multilayer coding scheme for efficient layer-wise decoding.
- Implemented simulated annealing to optimize system parameters.
- Modelled the quantum key distribution channel as a mixture of two Gaussians and a uniform distribution using probability, linear algebra, and machine learning.
- Wrote Bash scripts to run simulations on the UCLA Hoffman supercomputer and contributed as second author to a journal manuscript.
Skills: Python · C++ · Bash · Probability · Linear Algebra · Error-Correction Coding
UCLA · Electrical and Computer Engineering Department
Graduate Teaching Assistant
Los Angeles, California
- Served as a graduate teaching assistant for five quarters across Probability & Statistics, Data Science, Data Storage Systems, and Electromagnetic Waves.
- Led problem-solving discussion sections and office hours, and created and graded homework, examinations, and projects.
- View sample teaching evaluations.
Skills: Python · MATLAB · LaTeX
UCLA · Department of Radiation Oncology
Graduate Student Researcher
Los Angeles, California
- Worked on MRI image registration and generated a demonstration dataset to train LAPNet to estimate motion-field vectors between successive MRI images.
- Used Local All-Pass (LAP) estimation to obtain motion-field vectors between MRI slices for training.
Industry Experience
Advanced Micro Devices (AMD)
Performance Engineering Intern
Austin, Texas
- Benchmarked third-generation AMD EPYC Milan processors for machine-learning workloads using TPCx-AI and the MLPerf DLRM recommendation inference benchmark.
- Improved TPCx-AI benchmark performance by 10% through system and BIOS tuning while reducing hardware cost without sacrificing performance.
- Distributed the DLRM workload across 64 processor cores for greater efficiency.
- Completed the AMD EPYC certification covering data centers, AMD CPUs and GPUs, and EPYC architecture.
Skills: Python · Bash · Jira · GitHub
Perfios Software Solutions Private Limited
Data Science and Machine Learning Intern
Bangalore, India
- Worked on automatic table-structure recognition for bank statements using deep learning to convert images into machine-readable formats such as CSV.
- Developed a sampling algorithm for efficient graph neural network training on bank statements and built an inference pipeline for new image samples.
- Combined Tesseract OCR language features with CNN image features as graph neural network inputs.
- Measured model performance using precision and accuracy and presented the findings.
Skills: Python · TensorFlow · NumPy · Deep Learning