Hi there! I'm Seok Kyu Han, a passionate individual focused on integrating AI into the manufacturing industry. 🏭 My vision is to leverage predictive modeling and anomaly detection to drive smarter processes and achieve greater efficiency.

Data analysis and visualization are my forte. 📊 I thrive on uncovering hidden patterns in complex datasets and using them to solve real-world problems. I'm always eager to explore new visualization tools and share my findings on my tech blog. 🌱 My ultimate goal is to drive the intelligent transformation of South Korea's manufacturing sector through AI. Feel free to check out my repositories and reach out if you'd like to collaborate! 💻

• Developed predictive and anomaly detection models.
• Built web services for real-time monitoring.
• Specialized in optimizing manufacturing processes and energy savings.

1. MLFlowHub: Enterprise MLOps Platform (Prototype)

   • Duration: November 2024 - December 2024
   • Achievements: Delivered an MLOps platform prototype within 1.5 months, reducing the initial 3-month timeline.
   • Role:
     • Developed project, dataset, and experiment management features.
     • Integrated MLflow for model tracking and performance analysis.
     • Implemented JWT-based security and streamlined API interactions.
   • Technologies Used: Python, FastAPI, HTML, CSS, JavaScript, MySQL, Docker, MLflow, SQLAlchemy

2. Coke Oven Temperature Prediction Virtual Sensor System(Rollout)

   • Duration: March 2024 - August 2024
   • Achievements: Built prediction systems for 386 individual temperature targets with an MAE of 5.07℃.
   • Role:
     • Designed and implemented the system architecture.
     • Developed real-time anomaly detection algorithms.
     • Improved UI/UX for operator usability, including trust-level alerts.
     • Maintained system performance through quarterly updates.
   • Technologies Used: Python, Numpy, Scikit-learn, Flask, MySQL, TIBERO DB, JavaScript

3. Large-scale Rolling Mill Heating Furnace Virtual Sensor System Development

   • Duration: May 2023 - April 2024
   • Achievements: Transitioned heat calculation from yearly to real-time (1-minute intervals).
   • Role: Developed heat calculation and efficiency models, improved reliability of furnace operation.
   • Technologies Used: Python, Numpy, Pandas, SQL, Scikit-learn

4. Coke Oven Temperature Prediction Virtual Sensor System Development

   • Duration: June 2023 - December 2023
   • Achievements: Maintained prediction error within 7.36°C, expanded prediction variables from 1 to 71.
   • Role: Developed prediction algorithms, real-time monitoring system.
   • Technologies Used: Numpy, Scikit-learn, Flask, MySQL, Javascript

5. RTO Operation Stabilization in Chemical Plants

   • Duration: May 2022 - December 2022
   • Achievements: Achieved prediction relative errors of 2.7% and 5.4%.
   • Role: Developed VOC concentration prediction models, implemented real-time monitoring systems.
   • Technologies Used: Xgboost, LGBM, PCA, PLS, Flask, MySQL, JavaScript

6. Development of a Virtual Sensor System for Heating Furnace Exhaust Gas

   • Duration: June 2021 - December 2021
   • Achievements: Achieved over 85% accuracy in O2 and CO analysis.
   • Role: Developed non-linear prediction models, fault diagnosis system for sensors.
   • Technologies Used: PCA, PLS, NLPLS, A-ANN, TensorFlow, Flask, MySQL, JavaScript

• Hongik University, Seoul | Bachelor's in Chemical Engineering | GPA: 3.79/4.5
  • Research: Patent analysis and strategy for hydrogen storage systems

• ADsP (Data Analysis Semi-Professional) | March 2023
• SQL Developer (SQLD) | April 2023

• Grand Prize at Carbon Neutral Digital Industrial Innovation Big Data Platform Data Challenge | November 2022

• Flue Gas Analyzer Monitoring System Using Virtual Sensor Dualization for Reheating Furnace

  • Application Number/Date: 1020220150884 (2022.11.11)
  • This patent is related to the project "Development of a Virtual Sensor System for Heating Furnace Exhaust Gas."

• Smart Performance Monitoring System Using Real-time Efficiency Control Limit for Industrial Boilers

  • Application Number/Date: 1020210133362 (2021.10.07)
  • This patent is separate from other projects.

Feel free to explore my repositories and connect with me for any collaborative opportunities!