OpalAi is at the forefront of developing cutting-edge, award-winning technologies for creating digital twins of physical spaces, offering an experience akin to capturing video with a smartphone. Our work encompasses the development of innovative technologies as well as the adaptation of proven solutions originally designed for Mars and space exploration, now repurposed for terrestrial applications. Our mission is to revolutionize the mapping of indoor spaces, enabling seamless navigation within large facilities and indoor environments through the simple use of a smartphone. By bridging the gap between outer space innovation and everyday applications on Earth, we’re paving the way for the future of digital mapping and spatial computing.

OpalAi is a technology startup that develops automated digital twin reconstruction solutions using iOS app solutions and AI-enabled cloud processing.

Technical AI Mentor for NASA SBIR/STTR, Lead for Integrating novel optimization methods into Deep Space Network (DSN) scheduling platform, mentored students to advance careers in autonomous robotic systems. Mentored dozen students, two of my graduate interns participated in DARPA SubT challenge and they achieved astonishing results. Developed-derivate free optimization frameworks and autoencoder anomaly detection methods for applications from cybersecurity applications to data accountability of Mars rovers.

Developing advanced machine learning algorithms for the next generation of autonomous vehicles. During summer 2019, I am managing a group of 15 interns to build novel small satellites. I am working to automate the Deep Space Antenna using novel ML methods. Also, I am a leading ML and guidance for swarm autonomy.

Developing algorithms for swarm spacecraft autonomy and drones. Some of our work was demonstrated at the Caltech's Center for Autonomous Systems and Technology (CAST) opening ceremony.

At UCSD Flow Control Lab, we are developing a novel global optimization algorithm to deal with those optimization problems, which have both nonconvex and computationally expensive objective and constrained functions.

I was working on satellite and sky imagery solar power forecasting for a large array of solar photovoltaic (PV) cells.

I developed a derivative-free optimization framework (Surrogate Managment Framework) for expensive and time-consuming cardiovascular simulations.

We were working on uncertainty quantification (UQ) of solar power forecasting models using numerical weather prediction (NWP) for Virtual Power Plants (VPPs).