Ayush Kumar

ML engineer contributing to high-impact open source projects including Google Flax, Keras, and PyMC. Delivers production-grade features, bug fixes, and API improvements across core neural network and probabilistic computing libraries.

Open Source Contributor - PyMC (PyTensor), Google (Flax), Keras - Remote

(2025-01)

• Engineered Grouped Query Attention (GQA) support for Google Flax's MultiHeadAttention, optimizing memory bandwidth across core NN layers
• Resolved static shape inference bugs in PyTensor's pt.linalg.kron, adding regression test suites and expanding JAX backend support for BatchedDot
• Standardized Conv and LinearGeneral functional APIs across Flax and Keras, patching type-hint regressions and improving core documentation

B.Tech - Computer Engineering in Artificial Intelligence and Data Science - Ramrao Adik Institute of Technology, D. Y. Patil Deemed to be University (2023-07)