Kanishk Awadhiya

I’m a Quantitative Researcher with a Master’s in Artificial Intelligence from IIT Delhi, where I focus on high-performance computing and the optimization of large language models alongside complex mathematical modeling. My work bridges theory and systems: energy-based approaches, associative memory, and chaotic attractor dynamics.

I’ve authored multiple ICLR workshop papers and published research on themes like energy minimization, latent memory and associative memory mechanisms, and spectral bias in chaotic attractors. I’m comfortable turning research ideas into implementable workflows, especially when performance and numerical stability matter.

In my current Stealth Startup role, I develop low-latency trading algorithms using machine learning and optimize model inference pipelines using CUDA and vLLM. I aim to create solutions that are both research-grounded and operationally efficient for real-time use.

I also supported learning as a Teaching Assistant for machine learning and related advanced topics, reinforcing my communication and structured thinking. Whether through papers, prototypes, or performance tuning, I bring a rigorous, experiment-driven mindset to building intelligent systems.