Targeting production-grade robotic deployments for other enterprises by building custom action critique algorithms for frontier VLA models and data training pipelines with failure-prevention mechanisms for CI/CD development.

Developed quantized VLA architectures in PyTorch and embedded firmware interfaces in C, achieving 50% memory reduction and lower-latency inference for deployment on edge robotic platforms with limited compute and memory.

Engineered Python automation scripts to bridge legacy EHR systems with ML pipelines, reducing data processing latency by 95%, ultimately leading to better HIPAA end-to-end compliance for our hospital clientele.