Navjot Kukreja is a software engineer specialising in machine learning platforms and high-performance computing, with 10 years of experience building infrastructure for large-scale ML and scientific workloads. He combines academic rigor from a PhD in Computational Science with hands-on MLOps and platform engineering—having implemented production data pipelines with Metaflow, Kubernetes and Terraform and shipped PyTorch and scikit-learn models for time-series forecasting in energy. Navjot has a strong background in DSLs and compiler-level code generation, contributing to Devito by building the generator that turns SymPy-based stencils into efficient C kernels for wave propagation. His career spans academia (Assistant Professor and visiting researcher roles) and industry, including founding a stealth startup and leading platform work at Wayve, reflecting an ability to move ideas from research to production. Based in London, he brings a rare mix of HPC performance tuning and practical MLOps, often focusing on the low-level, non-obvious bottlenecks that make large-scale ML systems actually run.
10 years of coding experience
11 years of employment as a software developer
BITS Pilani, Birla Institute of Technology and Science
Doctor of Philosophy - PhD Computational Science, Doctor of Philosophy - PhD Computational Science at Imperial College London
Master’s Degree Operational Research, Master’s Degree Operational Research at University of Southampton
DSL and compiler framework for automated finite-differences and stencil computation
Role in this project:
Back-end Developer
Contributions:13 reviews, 622 commits, 130 PRs in 6 years 1 month
Contributions summary:Navjot is primarily focused on the Devito codebase's code generation and compilation features. They have built the `generator` class and related methods for handling and compiling C code stencils, the foundation of Devito. Their work involves building the interface from the Python code and SymPy based stencil definitions into executable C code and testing its functionality. This includes creating and testing different types of parameter handling to ensure the operators created can properly compute wave propogation stencils in 2D and 3D.
Python library to manage checkpointing for adjoints
Contributions:8 releases, 22 reviews, 158 commits in 5 years
python-librarypython
Find and Hire Top DevelopersWe’ve analyzed the programming source code of over 60 million software developers on GitHub and scored them by 50,000 skills. Sign-up on Prog,AI to search for software developers.