Clay Connors, Computer Science Developer in Cary, NC, United States
Clay Connors

Computer Science Developer in Cary, NC, United States

Member since May 9, 2022
Clay has extensive experience with data-oriented software design and development, having built a wide range of software to exact and open-ended customer specifications. He has also published research using deep neural networks for land use classification and change detection in high-resolution aerial imagery, later leading a team to develop a machine learning system that detects defects in solar panels from drone imagery.
Clay is now available for hire

Portfolio

  • Vadum
    Artificial Intelligence (AI), Software, Computer Science, Python...
  • Bamboo Mobile Health
    Java, Android, Mobile, DreamFactory, PHP, C, ARM Embedded, Amazon Alexa, Cron...

Experience

Location

Cary, NC, United States

Availability

Part-time

Preferred Environment

Linux, PyCharm, Conda, Git

The most amazing...

...public thing I've developed is an AI system for detecting physical changes in almost entirely unlabeled satellite imagery.

Employment

  • Project Engineer

    2019 - 2021
    Vadum
    • Wrote technical content of proposals in the electronic warfare space and implemented software and machine learning systems for accepted submissions.
    • Evaluated utility of classifiers and regression techniques for detecting varied kinds of interference.
    • Worked on a software system for fast approximate matching of time series data. Built the initial prototype of the system into a usable product and expanded the algorithm to include new methods that increase the system's speed and accuracy.
    • Expanded a system for secure computing with neural networks to work on full-scale, competitive deep neural networks with various layer types rather than only on fully-connected or convolutional layers.
    • Communicated with customers during development to give results and receive specifications. Built software to requirements and prepared demo-ready versions to display to customers at predetermined dates.
    Technologies: Artificial Intelligence (AI), Software, Computer Science, Python, Machine Learning, Deep Learning
  • Software Engineer

    2014 - 2017
    Bamboo Mobile Health
    • Assessed an incomplete build of an embedded wearable device, discovered all unimplemented capabilities and implemented them to bring the product to a demo-ready state.
    • Produced an Amazon Alexa skill to allow patients to verbally record medication events. The system vastly increases ease-of-use of patient self-reporting by being touch-free and effortless to use.
    • Developed an app to log patient symptoms for analysis alongside data from company servers. Allowed rendering a patient's symptom history through the app for physician diagnostic purposes.
    • Built a system to pull patient Fitbit data from third-party servers to company servers for analysis.
    Technologies: Java, Android, Mobile, DreamFactory, PHP, C, ARM Embedded, Amazon Alexa, Cron, Amazon Web Services (AWS)

Experience

  • Semi-supervised Change Detection in Very High Resolution Imagery
    https://github.com/c-connors/variational-change-detection

    I was given the task of leveraging recent developments in machine learning to provide highly accurate change detection in very high resolution (around one meter spatial resolution) bi-temporal aerial imagery.

    To allow using the object recognition capabilities of convolutional neural networks while avoiding high sample complexity, I took a semi-supervised approach. I trained a Variational Autoencoders (VAE), which optimizes an information-theoretic bound on latent information and allows training on a mixture of labeled and unlabeled data.

    I also modified the VAE to allow categorical and temporally-dependent latent variables in addition to the usual independent continuous latent variables seen in VAEs. This means that the neural network could learn, for example, that it was more likely for vegetation to transform into a building than the other way around. Areas with poor lighting in one-time steps could then use information from different time steps to decide on the class label.

    The quality of the predictions was highly satisfactory, considering the low labeling requirements. The paper was accepted to IGARSS 2017, where I presented the results and discussed potential future improvements to be made.

Skills

  • Languages

    Python, C, C++, Java, SQL, PHP
  • Other

    Artificial Intelligence (AI), Software, Deep Learning, Computer Science, Convolutional Neural Networks, Information Theory, Computer Engineering, Machine Learning, Geospatial Analytics, ARM Embedded, Conda
  • Libraries/APIs

    TensorFlow, Node.js, React, OpenCV, DreamFactory
  • Platforms

    Mobile, Android, Kubernetes, Amazon Alexa, Amazon Web Services (AWS), Linux
  • Storage

    Databases, MongoDB
  • Tools

    GIS, Cron, PyCharm, Git

Education

  • Master's Degree in Computer Science and Engineering
    2017 - 2019
    University of Michigan - Ann Arbor, MI
  • Bachelor's Degree in Computer Engineering
    2013 - 2017
    North Carolina State University - Raleigh, NC

To view more profiles

Join Toptal
Share it with others