Data Scientist Lead
Resume:
Henok Ghebrechristos, PhD
Lead Data Scientist, AI R&D Lead, AI Solution Architect
Contact
Meticulous and performance-driven Lead Data Scientist with 12+ years of experience in driving AI initiatives, integrating technological solutions, and directing & mentoring high-performing teams. Highly proficient in spearheading cybersecurity & data protection enhancements, AI-driven Software as a Medical Device (SaMD) delivery, maintaining strong vendor partnerships, and overseeing product development lifecycles. Adept at designing scalable architecture & platforms, AI software frameworks, and maintaining compliance with regulatory standards such as HIPAA. Specializes in computer vision and deep learning.
*****.**************@********.***
LinkedIn GitHub
Nashville, Tennesse
Education
PhD, Computer Science, and Information Systems, University of Colorado -Denver
Thesis - Exploiting Human Learning and Adaptation to Enhance Deep Learning: A Foundational Framework, Metrics and Experiments.
In progress, defending Oct 2024
Masters, Computer Science, 2014, University of Colorado-Denver
BS, Physics, 2010
University of Colorado – Boulder
Experience
AI SOLUTION ARCHITECT, LEAD DATA SCIENTIST COSTCO WHOLESALE, INC
OCT 2022 – CURRENT
Overseeing the implementation and adoption of AI technologies for enhancing various business domains.
Designed and implemented a hybrid AI solution architecture leveraging cloud services and on-prem compute/data warehouse.
Managing vendor partnerships to ensure the acquisition of high-performance Azure resources and on-prem computing equipment for AI/ML, GenAI, and RAG architecture development and deployment.
Leading and mentoring a team in deploying cutting-edge AI technologies for data protection and cybersecurity within a hybrid cloud architecture.
Architected and deployed a scalable platform, reducing data processing time by 90% and integrating custom Python transform engines and Azure Data Factory.
Enable a team of data scientists in developing custom AI solutions, frameworks, and products, for computer vision and cybersecurity.
Developed an end-to-end data ingest framework using Azure Data Factory, Azure Data Lake, SQLAlchemy, and pandas to transform and enrich data for operational analytics and AI workloads.
Leveraged MLflow, TensorFlow, and Keras for model performance monitoring and operational analytics, enhancing the accuracy and efficiency of AI workloads.
Increased information security effectiveness by 50% through the implementation of AI-powered threat detection systems.
Deployed a proprietary data labeling platform for computer vision tasks, enabling efficient annotation and training of warehouse-specific models.
Developed and deployed computer vision models to automatically detect defects in products to ensure quality.
Developed computer vision models for real-time inventory management, using deep learning object detection and tracking algorithms to identify and count items on shelves, track inventory movement, monitor stock levels, and alert staff when replenishment is needed.
DEEP LEARNING & COMPUTER VISION R&D LEAD, IMIDEX, INC
MARCH 2021 - 2023
Spearheaded the development of VisiRad XR, achieving FDA 510(k) clearance within a record timeframe of 13 months.
Assembled and led a high-performing R&D team, resulting in the successful integration of AI algorithms for lung nodule detection in VisiRad XR.
Oversaw the architectural design of a scalable cloud solution on GCP, successfully integrating AI algorithms for precise lung nodule detection.
Adopted engineering standards to ensure coordination and collaboration across the team.
Built a solution prototype to streamline data management and expedite data science and deep learning R&D.
Developed and released API package VisiRad.Cloud, interfacing with cloud infrastructures to streamline data movement.
Created VisiRad.Platform, a deep learning package specialized for radiology, supporting tasks such as object detection, segmentation and classification.
Resolved issues with the data curation platform, published the initial proprietary dataset, and established a consistent pipeline from training to validation.
Performed a proof-of-concept study demonstrating the solution's effectiveness on a standalone dataset.
Released VisiRad alpha, marking a significant milestone in the project.
SENIOR COMPUTER VISION ENGINEER, BOULDER IMAGING, INC
NOV 2012 - 2021
Use machine learning and statistical modeling techniques to develop and evaluate algorithms to improve performance, quality, data management and accuracy.
Leverage expertise in modern technologies, including TensorFlow, Keras, OpenCV, and C++, to develop solutions that drive growth, while also supporting existing/legacy applications.
Work in technical teams to develop, deploy, and apply deep learning-based computer vision systems.
Developed viAi (Vision Analytics and Intelligence) platform to streamline model development, validation, and deployment.
Engineered high-precision classifiers leveraging deep learning, CNN, leading to a 50% reduction in false positives for bird species identification.
Developed an automated dollar bill inspection system using Tesseract OCR and LSTMs.
Developed an AI system for wind farm eagle protection and integrating convolutional models (TensorFlow/Keras). This system has significantly reduced eagle deaths (as documented in Whitehouse testimony).
How New Technology Is Making Wind Farms Safer for Birds Audubon
Expertise
AI R&D Lead, AI Solution and Big Data Architecture
Deep Learning
AI ML Framework Design & Development, SDLC, CI\CD,
Leadership, Strategy,
Business Requirements,
Conference Publications, Presentation, & Teaching.
Projects / Achievements
VisiRadXR – Led the development of VisiRad XR, an AI-powered radiology platform. Developed and released critical components such as the API package VisiRad.Cloud and a deep learning package VisiRad.Platform. Achieved FDA 510(k) clearance within 13 months.
FDA Clearance News
IdentiFlight - Developed an AI system for wind farm eagle protection, integrating convolutional models to significantly reduce eagle deaths. Engineered high-precision classifiers leveraging deep learning, resulting in a 50% reduction in false positives for bird species identification.
Impact of the eagle protection surveillance system at wind farms: IdentiFlight System Impact
Languages, Frameworks, Tools
Programming Languages: Python, C++, C#, SQL
Frameworks: TensorFlow, Keras, PyTorch
Public Cloud: GCP, Azure
Big Data Technologies: PySpark, Databricks
Containerization: Docker, Kubernetes
Additional Skills: Databases, Modeling Tools, GenAI, MLOps, .NET, ASP.NET
Publications
For a comprehensive list of my scholarly contributions, please visit my Google Scholar profile: Google Scholar Profile
Contact this candidate