Devansh Bisla

Devansh Bisla

Senior Research Engineer (AI/ML) at NVIDIA

👤 About

Senior AI Research Engineer with a strong foundation in optimization methods, model architectures, and deep learning frameworks. Adept at converting research ideas into scalable AI applications.

🛠️ Skills

Languages: Python, C, C++, SQL
ML: PyTorch, JAX, TensorFlow
Cloud: GCP, AWS, AzureML, Docker
Inference: ONNX, TensorRT, vLLM

📰 Updates

  • Jan 2023 Our paper "ERASE-Net: Efficient Segmentation Networks for Automotive Radar Signals" was accepted in ICRA 2023.
  • July 2022 Joined NVIDIA, New Jersey full time as Senior Research Engineer (AI/ML).
  • April 2022 Successfully defended my dissertation thesis titled "Understanding Generalization in Deep Learning: An Empirical Approach".
  • Jan 2022 Our paper "Low-Pass Filtering SGD for Recovering Flat Optima in the Deep Learning Optimization Landscape" was accepted to AISTATS 2022.
  • May 2021 Excited to join Microsoft, Seattle as Data Scientist Intern.
  • Jan 2021 Our paper "A Theoretical-Empirical Approach to Estimating Sample Complexity of DNNs" was accepted to CVPR workshop track.