Applied Machine Learning Tech Lead with extensive experience launching innovative, cross-functional projects across video generation and editing (YouTube) and computational photography (Google Research). Currently working on AI for Science at Google DeepMind. Stanford EE PhD on optical and computational tools for medical imaging
Founding technical lead for the Self-Improving Meta Agent team in Google DeepMind's Science and Strategic Initiatives (SSI) organization. Accelerating scientific discovery and enterprise adoption by improving the usability, reliability, and quality of scientific models and agents, including AlphaEvolve, AlphaGenome, and more.
Technical lead for the Advanced Capabilities team in YouTube Shorts (applied ML, ~30 people; 7/2025 – 9/2026) and Research Scientist, Tech Lead for video generation in the Capabilities team (6/2023 – 7/2025). Collaborated with Google DeepMind on developing video generation models (evolving from Phenaki to Veo and Omni) to empower creators on the YouTube platform, scaling impact to >1.5M uploads per day. Notable projects include Dream Screen, AI Playground, AI Remix, and Conversational Editing.
Partnered with Google DeepMind to develop and launch Gemini Omni in YouTube Shorts, powering AI Remix and Dreamscreen.
Generative creation tools for YouTube Shorts.
Tech lead, manager (6/2020 – 6/2023), and developer of computational photography and computer vision algorithms for the Pixel phone and Google Photos. Core contributor to Night Sight (Pixel phone, since 2018), development of sky effects: Sky optimization for astrophotography (Pixel 2019) and Sky Palette transfer (2020), TL and IC for monocular depth (Cinematic Photos 2020, Bokeh 2022), Co-TL of Magic Eraser (Google Photos, since 2021).
Automatic distraction detection, instance segmentation, and on-device generative image inpainting for removing photobombers, power lines, and unwanted objects in Google Photos.
SIGGRAPH Asia 2019 paper and flagship Pixel camera feature introducing burst alignment, motion metering, and learning-based white balance for handheld photography down to 0.3 lux.
Long-exposure multi-frame sky segmentation, dark-current noise suppression, and star-trail prevention enabling mobile phones to photograph the Milky Way.
Research and development of computational imaging algorithms in the camera software group.
Dissertation: Optical and Computational Tools for Medical Imaging, advised by Prof. Adam de la Zerda; First lab member; developed optical design, signal processing, and computer vision tools for optical coherence tomography (OCT) and hyperspectral microscopy, publishing 4 first-author journal papers.
Stanford Ignite (Graduate School of Business, Jun – Jul 2016): Completed a certificate program providing business fundamentals for entrepreneurship, working with a small team to develop a business model and commercialization plan for one of my PhD medical imaging projects.
Eliminated coherent speckle noise in Optical Coherence Tomography using a diffusing element, revealing cellular-scale structures in living tissue and the human retina.
Engineered large gold nanorods and spectral signal reconstruction for non-invasive 3D visualization of lymphatic drainage, valve function, and tumor microvasculature in vivo.
Adaptive detection and machine learning classification (SVM/k-means) of single plasmonic nanoparticles inside scattering biological tissues.
Built a real-time 60 FPS volumetric 3D OCT renderer for Oculus Rift (EE 267) and an Android face detection + OpenGL dynamic illumination blending app (EE 368).
Closely led a small team of engineers conducting R&D in computer vision, working on a product to assist in cardiac navigation procedures by creating a 3D model of the heart from two X-Ray projections.
Performed research, design, and implementation of computational imaging algorithms and created tools for image quality evaluation for cellphone cameras. Primarily developed an algorithm for Extended Depth of Field (EDoF), which was incorporated in numerous cellphones by Nokia and Samsung.
Computational imaging algorithms and optical-digital co-design deployed across commercial Nokia and Samsung mobile phone cameras.
Advised by Dr. Yael Hanein. Research on dissipative particle dynamics model of carbon nanotubes, contributing to the open-source particle dynamics simulation project SYMPLER.
Nanoscale particle dynamics simulation of carbon nanotube self-assembly and mechanical behavior in SYMPLER.
Led a team at the core of major multi-disciplinary projects in the field of electro-optics. In charge of building and testing complex systems with tight deadlines, alongside R&D in electro-optics and embedded software development and system design.
Student engineer. Design of high-frequency microchips, electromagnetic simulations of distributed systems, and modeling of components.
Majored in communication, wave theory, and electro-optics. Projects on numerical analysis of electromagnetic waves and high-speed fiber optic lasers.
Head Teaching Assistant for two years during my PhD at Stanford (with Prof. Gordon Wetzstein) and invited annual guest lecturer on mobile computational photography.
Speaker at GHC 2019 (with Yael Pritch) on AI for computational photography on Google Pixel, covering single-camera Portrait Mode and handheld low-light photography with Night Sight.
Featured talk on building and launching generative AI creation tools for video and photography.
Full recording of my doctoral defense at Stanford University covering Speckle-Modulating OCT, MOZART gold nanorod contrast agents, and hyperspectral dark-field microscopy.
A lighthearted throwback video from my Stanford PhD days in the lab.
A determined mid-packer, exploring trails and distances from 5Ks to 100Ks and hopefully one day 100 miles.
Painting, playing with shapes and color.
