Show HN: Gradient-Free ML Algorithm (Available for Contract Work)

Hi HN,

I'm Allen Townsend, an independent ML researcher, and I've developed a novel gradient-free machine learning algorithm that enables faster training and inference than traditional gradient-based approaches, especially in scenarios where:

Gradients are hard or impossible to compute

Simulation or black-box models dominate

Speed and stability are critical

This method avoids backpropagation entirely, yet achieves competitive results in optimization-heavy tasks. It's well-suited for reinforcement learning, simulation-based systems, and edge deployment use cases.

I'm currently looking for contract work to help teams develop custom models using this approach. If you have a tough modeling or optimization problem and want to explore this new direction, I’d love to collaborate.

You can reach me at [email protected].

Happy to provide examples, benchmarks, or run a quick pilot project. Thanks for reading!

— Allen Townsend