As a principal engineer at NVIDIA focusing on deep learning and generative AI, I play a key role in driving the development of advanced algorithms and system architectures. My work covers the entire process from research innovation to practical application, focusing on improving the performance and efficiency of AI models.

Worked on designing cloud-based AI services, focusing on privacy-preserving Large Language Models (LLMs) and DeepFace detection.

Also worked as a driving force behind the development of Intel's Homomorphic Encryption (HE) ecosystem, I have led a research and development team in building the fundamental components for HE solutions. The efforts have resulted in the creation of 5 open-source software libraries to accelerate HE, facilitate adoption, benchmark performance, and provide API primitives.

As a lead architect in AI, Machine Learning, and Media Analytics solutions, I have successfully co-designed products with major customers using a range of Intel offerings, including Intel GPU, FPGA, VPU/Movidius, Nervana NNP-I, media components, and speech recognition accelerators. My expertise in these areas has allowed me to drive innovation and deliver impactful results for our customers.

Led a path-finding team that explored new applications of Intel GPUs to accelerate Deep Learning, Computer Vision, Augmented Reality, and Big Data Analytics. One of creators of clCaffe, an open source extension of Caffe framework with OpenCL backend for Intel GPUs.

Implemented a mesh simplification algorithm that were able to reduce count of triangles that represent salt body or horizon surfaces by nearly 60% without degrading quality of the terrain information. Among other responsibilities, prototyped a visualization tool to render slices of seismic data using 3D volumetric rendering.