Jul 2018 – Present · 7 yrs 10 mos · San Francisco Bay Area · On-site

• Currently, leading teams in various domains - manufacturing of semiconductors, inspection and metrology, EDA analysis, agriculture + AI, aerospace.
• Previously in medical instruments, autonomous vehicles, flood detection, space, meteor tracking.
• Responsible for end-to-end training and deployment of computer vision optimizations in medical instruments and drug discovery NVIDIA partners globally.
• Deployed NVIDIA Autonomous Vehicle software at scale - led multiple bring-ups of the entire software stack on thousands of NVIDIA GPUs
• Led Nvidia’s Flood and Disaster Relief Strategy team - an initiative with UN to produce an educational course on deep learning for flood detection.
• Partner Advisor for NASA Frontier Development Lab (AI Accelarator)

May 2017 – Present · 9 yrs · Mountain View, California, USA

• I advise and mentor NASA FDL researchers to develop machine learning and deep learning tools to impact a variety of real-world challenges.
• Highlights include:
• (1) creating a meteor detector for NASA CAMS that keeps eyes on the skies so scientists can sleep at night, and is one of the most used tools by astronomers that detected the highest number of meteors in a single night in NASA’s 63-year history: https://www.seti.org/cams. 10 new comets discovered, 20 parent bodies identified, and instrumental evidence of previously predicted C/1907 G1 (Grigg-Mellish) comet discovered.
• (2) developing a flood zone segmentation technique that can analyze 24,000 square miles in just 3 seconds; transformed into an educational course offered by United Nations Satellite Centre and deployed in 5 countries: https://siliconvalley.orange.com/en/news/rapid-disaster-detection-and-response-a-climate-action-collaboration/
• (3) enabling the Worldview Search pipeline that curates datasets from over 20 years of satellite imagery (petabytes) in hours instead of months (manually): https://www.earthdata.nasa.gov/learn/articles/spaceml-impact-blog
• (4) DAGGER model - developed a 30-minute forecasting method to predict Geomagnetically Induced Currents which form the backbone of modern electric grid and telecommunication satellites. https://www.nasa.gov/science-research/heliophysics/nasa-enabled-ai-predictions-may-give-time-to-prepare-for-solar-storms/ Work followed up by the SURYA model https://science.data.nasa.gov/features-events/inside-surya-solar-ai-model

Feb 2017 – Jul 2018 · 1 yr 5 mos · Palo Alto

• Acquired by NXP Semiconductors.
• As the first engineer and only Deep Learning Data Scientist in a 2-person startup (DeepVision, now called Kinara), I developed deep learning models for resource-constrained electronic devices.
• Work ranged from research to prototyping to engineering, and deploying deep learning based high-performance data science models, including computer vision algorithms at scale on servers as well as on everyday lightweight, power-aware, resource-constrained edge devices like self-driving cars, smart security cameras, and robots. Also led the deployment of these algorithms on Deep Vision’s proprietary processor chip targeting these markets.

Jan 2017 – Oct 2019 · 2 yrs 9 mos

• 600 pages of Real-World AI & Computer Vision Projects Using Python, Keras & TensorFlow
• https://www.oreilly.com/library/view/practical-deep-learning/9781492034858/
• Whether you’re a software engineer aspiring to enter the world of deep learning, a veteran data scientist, or a hobbyist with a simple dream of making the next viral AI app, you might have wondered where do I begin? This step-by-step guide teaches you how to build practical deep learning applications for the cloud, mobile, browser, and edge devices using a hands-on approach. Relying on years of industry experience transforming deep learning research into award-winning applications, we guide you through the process of converting an idea into something that people in the real world can use.
• Train, tune, and deploy computer vision models with Keras, TensorFlow, Core ML, and TensorFlow Lite
• Develop AI for a range of devices including Raspberry Pi, Jetson Nano, and Google Coral
• Explore fun projects such as Silicon Valley’s "Not Hotdog" app to image search engines and 40+ industry case studies
• Simulate an autonomous car in a video game environment and then build a real miniature version with reinforcement learning
• Use transfer learning to train models in minutes
• Discover 50+ practical tips on maximizing model accuracy and speed, debugging, data collection, avoiding bias, and scaling to millions of users
• Features luminaries including François Chollet, Jeremy Howard, Pete Warden, Anima Anandkumar.
• Referenced as recommended reading for CS663 Computer Vision course, Cal State East Bay, and Karlsruhe Institute of Technology, Germany for Edge-AI in Software and Sensor Applications course.
• Featured on Keras website as a learning resource: https://keras.io/getting_started/learning_resources/

May 2016 – Dec 2016 · 7 mos · Greater Pittsburgh Area

• Research focused on deep learning models with weak supervision: Utilizing supervision from visual questions asked about images. Spotlight presentation & poster at the IEEE Computer Vision and Pattern Recognition conference, 2017.
• Mentors: Olga Russakovsky, Abhinav Gupta

Jan 2016 – Dec 2016 · 11 mos · Greater Pittsburgh Area

• Open Advancement of Question Answering Consortium.
• Researched Deep Learning techniques for text question answering systems. Developed Question-Answering systems based on an ensemble of Deep Learning and Rule-based systems for Natural Language Processing Systems on the SQuAD dataset.
• Mentors: Eric Nyberg, Matthias Grabmiar

May 2016 – Present · 10 yrs

• Delivered 200+ keynotes across 50+ countries, including EuroPython, Deep Learning World, Strata, O'Reilly, TWIMLAI, WeAreDevelopers
• Served as judge for various competitions, including CES Innovation Awards.
• Academic Conference Reviewer: NeurIPS, CVPR, ICCV, AAAI, ECCV, WACV, ACCV, IEEE Big Data, GHC, WiML. Industrial: GTC, O'Reilly, Strata.
• Book reviewer for multiple publications, including the widely popular "Hands-On Machine Learning with Scikit-Learn, Keras, and TensorFlow" by Aurélien Géron
• Reviewer for multiple NASA programs, including Living With a Star https://lws.gsfc.nasa.gov/

Jun 2015 – Jan 2016 · 7 mos

• Leading contributor and supporter for Open Research.
• Member of the WOW! OLC, Open Leadership Cohort (https://science.mozilla.org/blog/wow-introducing-working-open-workshops-and-the-open-leaders-cohort)

May 2015 – Aug 2015 · 3 mos · Geneva Area, Switzerland

• Atom Smashing using Machine Learning and Deep Learning at CERN
• Used Apache Spark to streamline different predictive prototypes by gathering information from CMS data-services, ran predictive models and suggested which datasets will become popular over time achieving Dynamic Data Placement and efficient resource utilization. Then, evaluated quality of individual models, performed component analysis and chose best predictive model for new set of data. This included evaluation of Apache Spark as Analytics framework for CERN’s Big Data Analytics infrastructure.
• DOI: 10.31861
• Mentors: Tony Wildish, Valentin Y. Kuznetsov, Manuel Martin Marquez, Antonio Martin Marquez
• Won the CERN Webfest 2015 in Best Innovative Outreach