Course Projects
Deep Learning Systems: Algorithms and Implementation
Built a PyTorch-like deep learning framework from the ground up, with backend support of CPU/GPU containing standard functionalities such as automatic differentiation, optimizers, data loaders, loss functions, and all required modules to employ parametrized layers.
(Python, C++, CUDA)Designed and implemented a NeuralODE layer (operators and backends, ODE numerical solver, AD) within our Pytorch-like framework and demonstrated its capability of approximating ODE dynamics with NNs. (Python, C++, CUDA)
ML with Large Datasets
Conducted various analyses such as entity resolution and PCA and built ML pipelines on large datasets such as Million Song Dataset, light-sheet imaging, and Criteo 1TB click logs dataset. (Python, AWS EC2, PySpark)
Implemented multiple model compression techniques from scratch (network slimming, magnitude-based pruning, and …). (Python, Tensorflow)
Quantum Integer Programming
- Implemented a quantum computing MILP solver using D-wave quantum computer for job shop scheduling problem (Python)