Working on developing state of the art GPU Kernels, for modern systems (Nvidia Blackwell, AMD MI300). The kernels written power the latest models in the industry, including DeepSeek, GPT OSS and many more.

• Raised 150,000 from ERANYC (largest startup accelerator in New York City)

• Led & Developed an optimized AI inference platform. Utilized tools such as CUDA, NEON, and MKL, to write custom ML Kernels for various model architectures. Created Deployment infrastructure using AWS orchestration tools and Python. Developed custom agentic platforms utilizing the latest Generative Vision and Language models.

• Researched and developed systems to improve self-driving car accuracy using technologies like Deep

Audio Analysis, Docker, and TensorFlow.

• Trained in-house models using GANS and other computer vision models. Made a GPU accelerated image generator for hyperrealistic road-related images and an object Inpainting tool.

• Converted car views from fish-eyed lenses to bird's eye views using technologies like multi-layer

perceptrons, skip connections, and CNNs.

• optimized models for various edge devices including prototyped R/C cars. This involved utilizing custom GPU based kernels, and model export processes.

Developed Hardware optimized recommender (for gpus) systems for simpler algorithms such as Content Based Collaborative Filtering & Content Based Collaborative Filtering.

Utilized Pytorch to train Deep recommender Systems such as Neural Collaborative Filtering Models, and Matrix Factorization Models.