UC Santa Barbara

Machine Learning Research Intern

Jun 2024 - Aug 2024 · Santa Barbara, CA

  • Benchmarked CNNs, MLPs, and SNNs in PyTorch/SNNtorch to study accuracy-latency tradeoffs for audio classification.
  • Built MFCC ML pipelines with model profiling and optimization, reducing inference latency by 37% on 10K+ audio samples.
  • Evaluated model performance with accuracy, latency, and confusion-matrix analysis; presented findings at ICOPH Bangkok.

Summer research on the accuracy-latency tradeoff between CNNs, MLPs, and spiking neural networks for audio classification. Built MFCC pipelines and profiling tooling that cut inference latency 37% across 10K+ samples, and presented the findings at ICOPH Bangkok.

pytorchsnntorchaudioresearch