Career

Scale AI
May 2024 ~ August 2024

LLM Trainer

• •  LLM Response Evaluation and Enhancement for Fine-Tuning
    - Enhanced the fine-tuning pipeline by systematically evaluating and refining outputs generated by Large Language Models (LLMs), utilizing a rigorous approach to identify model weaknesses and implement optimized responses.

  • Applied deep linguistic analysis and coding expertise to improve LLM performance in native Korean language tasks, focusing on syntactic, semantic, and contextual accuracy.

Jeju Air (IATA: 7C, ICAO: JJA)
July 2018 ~ May 2024

Data and System Engineer
August 2022 ~ May 2024

• • Data Engineering and Machine Learning
    - Developed a centralized data lake using Apache Kafka for real-time ingestion of aircraft sensor data and Apache Spark for large-scale processing, enabling predictive flight operations
    - Implemented a Retrieval-Augmented Generation (RAG) framework by integrating Large Language Models (LLMs) with real-time data sources, allowing for enhanced contextual understanding and predictive analytics in maintenance decision-making

  • Cost analysis and Optimization
    - Conducted regression analyses based on the usage trend  over  5 years and suggested a better usage model
    - Achieved a 39% reduction* in datalink communication** expense and eliminated duplicated services
    - Awarded Model Worker of the Year out of 3,000+ employees (2023)
    * Approximately more than $661,000 a year
    ** Digital messages between aircraft and ground stations or other aircraft, known as ACARS  

  • Project Planning and Scheduling
    - Developed and monitored project timelines to ensure timely delivery of milestones, using tools such as Slack, Gantt charts (Excel)

  • Communication and Reporting
    - Established communication channels with stakeholders through regular and ad hoc meetings, emails, and reports
    
    Key Projects
    - Scheduling System Examination
    - Safety Management System (SMS) Project
    - Aeronautical Information System Project