Applied Scientist
CurrentMeenthttps://github.com/kc-ml2/meent- Developed 'Meent': code for optical simulations and optimization. -- https://github.com/kc-ml2/meent -- implemented on Numpy, JAX and PyTorch. -- Supports Automatic Differentiation (AD) which returns gradient of loss per each input variable. -- with the help of JAX pmap, Meent support parallelization(CPU and GPU) - Role in Meent -- As a physicist, have strong understanding on optics and mathematical analyses enough to fix and implement complicated formulations in the field of optics. -- As a software engineer, solely designed and developed Meent, which is a total solution to model a structure, simulate its optical characteristics and solve optimization problem based on gradient-based algorithms. -- As a project manager, adjust priority, make community and promote the product for researchers to participate or use by providing immediate support and guidance.- ML applications in semiconductor industry -- Participating research on metasurface design using by RL -- Developed image detection code running on Edge device(RPi4, Jetson) -- Investigating Physics-informed-ML- Tools -- Python, JAX, PyTorch, Tensorflow, Docker, MATLAB, Linux, GCP, HPC