- AI infrastructure for data centers- Developer tools for AI computing & help with AI strategy- Affordable GPUs

- Affordable and easy-to-use compute for AI- On-prem and datacenter cluster management- AI model deployment

Leading efforts in the space of GenAI powered content creation. Working on text-to-3D and image-to-3D problems.

Leading a team of software engineers and machine learning experts developing an in-house deep learning framework with the focus on network optimization for mobile platforms. It is powering over forty use cases like Animoji, Portrait Mode, ARKit and is used to optimize nearly all mission critical neural networks in Apple Vision Pro.

Leading a team of software engineers and computer vision experts to develop a 3D foot scanning solution for HP FitStation platform. It enables 3D foot scanning, foot measurement and shoe recommendation services.

• Developing algorithms for robotics navigation and vision using Python, C++, OpenCL and CUDA.• Developing algorithms for vision-based manipulation, writing control software for Baxter manipulator. Implemented learning algorithm trainable by demonstration to distinguish several types of objects and perform various actions based on the object type, location, sound input and other types of input.• Leading a team of five software engineers during spin-off mobile-manipulator project. We have built mobile manipulator platform with several cameras, implemented vision system to detect objects on the floor, implemented control software for the manipulator to pick objects from the floor and move them to the predefined location.• Prototyping sensors and robotic controllers, designing PCB boards with Eagle software, writing firmware for STM32 microcontrollers.

• Developing software for hand-held 3D scanners, including professional Artec scanners and consumer Kinect, Asus Xtion, etc.• Developing 3D scanning and geometry processing algorithms. Implemented GPU pipeline for 3D scanning similar to KinectFusion: registration of depth-maps using ICP, fusion of depth maps into sparse volumetric representation, extraction of isosurface with marching cubes algorithm.• Developing application logic of the 3D scanning software using C++ and Python programming languages.

• Developing user-mode DirectX video driver to allow playing games with stereoscopic devices: parallax barrier monitors, shutter glasses, etc. Developed fast memory allocator, designed simpler command buffer manager architecture for stereoscopic rendering, implemented logging system suitable for user-mode driver• Developing software for automatic user-interface testing with AutoIt.• Prototyping computer vision algorithms for converting regular videos to 3D videos.

• Developing graphics effects for computer game engine using OpenGL & GLSL. Developed several graphics effects: deferred shading lighting system, mass object rendering with pseudo-instancing and BVH culling, projected decals, post-effect system (with effects like ”Depth of Field”, ”Bloom”).• Developing game engine logic using C++ programming language.

• Developing graphics effects for 3d interactive presentations using OpenGL & GLSL. Implemented several algorithms: marching cubes algorithm for the visualization of volumetric data using geometry shaders, visualization of the ocean surface with reflections and refractions, fire and magma effects, HDR rendering with tone mapping.• Programming application logic using C++ programming language.• Maintaining CMake based build system.