Skip to content
View Cho-SungHyun's full-sized avatar
1 follower · 0 following

Block or report Cho-SungHyun

Report abuse

Contact GitHub support about this user’s behavior. Learn more about reporting abuse.

Report abuse
Cho-SungHyun/README.md

Hi there, I'm Seonghyun Cho 👋

Data Scientist | Machine Learning Engineer | AWS Certified ML Specialist

I am a Data Scientist with an MSc in Artificial Intelligence and a BSc in Computer Science. Currently, at Living Optics, I specialise in building production-ready ML solutions, focusing on Multimodal Fusion, MLOps, and Statistical Optimisation.

I bridge the gap between advanced R&D and commercial scalability, ensuring data integrity through rigorous unit testing and automated validation.

🛠 Technical Toolkit

  • Languages: Python (Expert), SQL (Expert), MATLAB, C
  • Frameworks: PyTorch, Scikit-learn, OpenCV, Pandas, NumPy
  • Specialties: Multimodal Learning, Late Fusion CNNs, Bayesian Optimisation (BOHB)
  • Cloud & DevOps: AWS (Certified ML Specialist), Git, Unit Testing, CI/CD Workflows

📚 Featured Research & Projects

Music Genre Classification via Late Fusion CNN

  • Implementation of the Late Fusion strategy using STFT, MLS, and MFCC spectral features.
  • Comparative analysis of multi-modal features for enhanced classification accuracy.
  • Published in IEEE Access, 2024.

Efficient Neural Network for CT Image Analysis

  • Research on channel compression techniques to accelerate medical image processing.
  • Focus on balancing computational efficiency with diagnostic precision.

📈 Certifications & Achievements

  • AWS Certified Machine Learning – Specialty (2024)
  • Lead Author, "A case study on the integration of a snapshot hyperspectral field-portable imager solving fruit quality assessment," SPIE, 2025.
  • Lead Author, "CNN-based Music Genre Classification using Multiple Spectral Features," IEEE Access, 2024.

📫 Connect with me:

LinkedIn | Email

Pinned Loading

  1. Music-Information-Retrieval-Genetic-Algorithm Music-Information-Retrieval-Genetic-Algorithm Public

    Evolutionary Feature Selection (Genetic Algorithm) for high-dimensional music data. Published in IEEE Access (2021).

    MATLAB 1

  2. An-Efficient-Neural-Network-based-on-Early-Compression-of-Sparse-CT-Slice-Images An-Efficient-Neural-Network-based-on-Early-Compression-of-Sparse-CT-Slice-Images Public

    High-efficiency CNN for medical imaging. Achieved ~3.5x actual inference speedup over ShuffleNet-v2. Published in IEEE Access.

  3. Music-Genre-Classification-using-Late-Fusion Music-Genre-Classification-using-Late-Fusion Public

    SOTA Music Genre Classification using Late Fusion CNN. Evaluated on 12 public datasets. Published in IEEE Access (2024).

  4. Music-Genre-Classification-Using-Multiple-Musical-Visual-Features Music-Genre-Classification-Using-Multiple-Musical-Visual-Features Public

    Multi-input DenseNet architecture with data augmentation (sliding window/overlap). Published in IEIE (2022).