Jihye Park

Software Engineer, Samsung Electronics

September 2023 – Present

Developed industrial AI services for semiconductor manufacturing across agentic AI, predictive AI, and production-ready AI pipelines.

• Designed and implemented agentic AI services that combine RAG-based information retrieval, domain knowledge-based query understanding, and meta data-based dynamic prompt engineering.
• Developed predictive and explainable AI applications for manufacturing data, including root-cause identification, anomaly detection, and BERT-based terminology-alias matching.
• Built and evaluated production-ready AI pipelines using Airflow, Kubernetes, and Docker for scalable processing and real-time service operation.

Backend Developer, Bankware Global

July 2016 – June 2018

Worked as a backend developer using Java, Linux, and Oracle.

• Worked as a member of the "Software Research Laboratory (소프트웨어 연구소 > 인프라 솔루션 팀 & 인터페이스 솔루션 실)."
• Developed backend process of the OK Bae-Jung Foundation scholarship website, communicating with frontend developers and UI/UX designers.

Market Researcher, Hankook Research

January 2016 – May 2016

Analyzed user-generated data mainly obtained from surveys by using MS Excel.

• Conducted group surveys and home visiting interviews in the Los Angeles, United States, to create a report on built-in home appliances.
• Conducted several interviews with experts in the digital signage field.

Seoul National University

Ph.D. in Industrial Engineering (공학박사) • August 2023

• Majored in Data Mining
• Advised by Professor Sungzoon Cho
• Dissertation: "Financial Risk Assessment Automation: Hot Topic Detection in Speeches, Sentiment Analysis of News Articles, and Spam Filtering on Twitter"
  (금융위험 평가 자동화: 연설문 핫토픽 탐지, 뉴스 감성 분석 및 트위터 스팸 필터링) [View]

Korea University

Bachelor of Business Administration • February 2016

• Internship at Linewalks (2014): Suggested ideas for a new service employing healthcare big data and designed marketing promotion materials.
• Internship at Hankook Tire Hungary Ltd. (2016): Worked with Hungarian colleagues dealing with EU taxes, communicating in English.

Lecturer (강사), Chung-Ang University, Department of Advertising and Public Relations, "Digital Analytics" Course

Spring 2023

중앙대학교 경영경제대학 광고홍보학과 전공선택 과목 "디지털 애널리틱스"
[Lecture Notes]

Question Structuring First: Execution Planning for Industrial AI Agent Systems

This paper proposes a question-structuring-first approach to execution planning in industrial AI agent systems. Instead of directly generating execution plans from user queries, the framework first structures the question into task categories, required outputs, referenced metrics, and argument policies to improve agent selection, execution stability, and latency.

Keywords: AI agent systems, question structuring, execution planning, agent selection, task decomposition, tool-augmented reasoning, industrial AI

Plan-Then-Write: Question-Guided Table-to-Report Generation with Large Language Models

This paper explores question-guided table-to-report generation with large language models. It proposes a plan-then-write framework that first identifies the user’s analytical intent, relevant table evidence, and report structure before generating faithful and informative reports.

Keywords: table-to-report generation, large language models, question-guided generation, evidence planning, numerical accuracy, self-verification

Not All Relevant Sentences Are Evidence: Direct Evidence Sentence Selection for Faithful RAG

This paper studies the distinction between relevant sentences and direct evidence sentences in Retrieval-Augmented Generation. It proposes a sentence-level evidence selection framework to improve answer faithfulness and reduce unsupported claims in domain-specific RAG systems.

Keywords: Retrieval-Augmented Generation, faithful generation, evidence sentence selection, hallucination reduction, unsupported claim detection, claim-evidence alignment

Automatic Construction of Direction-aware Sentiment Lexicon Using Direction-dependent Words [View]

Jihye Park, Hye Jin Lee, Sungzoon Cho* (2024), "Automatic Construction of Direction-aware Sentiment Lexicon Using Direction-dependent Words," Language Resources and Evaluation, Volume 59, pages 843–869, 25 May 2024.
● Language Resources and Evaluation (LRE) has been indexed in the SCIE as of May 2024.

Incorporation of Company-Related Factual Knowledge into Pre-trained Language Models for Stock-Related Spam Tweet Filtering [View]

Jihye Park, Sungzoon Cho* (2023), "Incorporation of Company-Related Factual Knowledge into Pre-trained Language Models for Stock-Related Spam Tweet Filtering," Expert Systems with Applications, Volume 234, 121021, 30 December 2023.
● Expert Systems with Applications (ESWA) has been indexed in the SCIE as of December 2023.
● ESWA has 2022 Impact Factor of 8.5, based on the Journal Citation Reports by Clarivate Analytics (released in June 2023).
[Data] [Code]

Building Knowledge Graphs from Technical Documents Using Named Entity Recognition and Edge Weight Updating Neural Network with Triplet Loss for Entity Normalization [View]

Sung Hwan Jeon, Hye Jin Lee, Jihye Park, Sungzoon Cho* (2023), "Building Knowledge Graphs from Technical Documents Using Named Entity Recognition and Edge Weight Updating Neural Network with Triplet Loss for Entity Normalization," Intelligent Data Analysis, Volume 28 No. 1, pages 331-355, 30 November 2023.
● Intelligent Data Analysis (IDA) has been indexed in the SCIE as of February 2024.

Hot Topic Detection in Central Bankers' Speeches [View]

Jihye Park, Hye Jin Lee, Sungzoon Cho* (2023), "Hot Topic Detection in Central Bankers' Speeches," Expert Systems with Applications, Volume 230, 120563, 15 November 2023.
● Expert Systems with Applications (ESWA) has been indexed in the SCIE as of November 2023.
● ESWA has 2021 Impact Factor of 8.665, based on the Journal Citation Reports by Clarivate Analytics (released in June 2022).
[Data] [Code]

가짜뉴스 문제 해결을 위한 범국민 SNS 캠페인 제안

윤서윤 (중앙대학교 광고홍보학과)†, 조영아 (중앙대학교 도시계획부동산학과)†, 최영현 (중앙대학교 프랑스어문학과)†, 박재홍 (중앙대학교 역사학과)†, 박지혜*
[2023 KOSAC "시민 문화 정착 프로젝트 기획" 공모전 출품]

Music-Circles: Can Music Be Represented With Numbers?

Seokgi Kim†, Jihye Park†, Kihong Seong†, Namwoo Cho†, Junho Min†, Hwajung Hong*
arXiv preprint arXiv:2102.13350, 2021.
[Media Release] [Paper] [Demo] [Video] [Code]

Teaching Assistant, NH Nonghyup Bank (NH농협은행) "4th Industrial Revolution Core Talent Cultivation" Course

Summer 2022, Summer 2023

Provided technical support for big data analysis program for NH Nonghyup Bank employees.

Lecturer, Korea Arts Management Service (예술경영지원센터) "Data Analysis" Course

Fall 2022

Presented lectures on MS Excel-based data analysis techniques.

Lecturer, Grand Korea Leisure (그랜드코리아레저) "Data Analysis" Course

Summer 2022

Presented video lectures on MS Excel-based data analysis techniques.

Lecturer, Kyungbok University (경복대학교) "Basics of Big Data Analysis" Course

Summer 2022

Presented video lectures on MS Excel-based data analysis techniques.

Teaching Assistant, MKYU "Big Data Analysis" Course

Spring 2022

Provided feedbacks on data-driven value creation proposals written by MKYU students.

Teaching Assistant, LG U+ (엘지유플러스) "Dream Big Data" Course

Fall 2021

Provided technical support for Digital Transformation (DX) training program for LG U+ employees.

Lecturer, SNU Language Education Institute (서울대학교 언어교육원)

November 2019 – February 2021

Presented lectures on Python programming and machine learning for government-invited Malaysian students.

Teaching Assistant, SNU Data Mining Camp / SNU Big Data AI Camp

February 2019 – August 2023

Assisted lectures on introduction to data mining for high school students.

주성분의 특징을 이용한 대량 반도체 웨이퍼의 광학스펙트럼 전 파장 데이터 비교분석 방법

특허등록번호 P2952371 (2026.04.09)

차원 축소 알고리즘과 군집 분석 평가 지표를 활용한 TTTM(Tool-to-Tool-Matching) 활동 관리 지수

특허출원번호 P20240116940 (2024.08.29)

차원 압축 도메인에서의 가이드라인을 이용한 Biplot 분석 방법

특허출원번호 P20240110061 (2024.08.16)

센서 관련 지식이 강화된 언어 모델을 활용한 자동 센서명 분류 방법

특허출원번호 P20240102576 (2024.08.01)

센서 관련 지식이 강화된 언어 모델을 활용한 센서명 자동 분류 방법

박지혜, 최민훈, 윤동운, 유하늘, 임재용, 최인수, 김수정, 김재혁, 정지수, 조성민, 금의석
Fall Conference of Korean Institute of Industrial Engineers (KIIE), Seoul, Korea, 2024.

Machine Learning 기반 반도체 공정/설비 불량 원인 분석 방법론 [View]

유하늘, 박현섭, 임재용, 김수정, 윤동운, 박지혜, 조성민, 금의석
Spring Conference of Korean Institute of Industrial Engineers (KIIE), Yeosu, Korea, 2024.

PacMAP을 이용한 반도체 제조 공정 데이터에서의 이상 탐지 [View]

유용재, 최진우, 박지혜, 조성준*, 유하늘, 박광균, 임재용, 박현섭, 윤동운, 김수정, 조성민, 금의석
Fall Conference of Korean Institute of Industrial Engineers (KIIE), UNIST, Korea, 2023.

Post-training Approach Using Form 10-K Filings to Inject Financial Domain-specific Factual Knowledge into Pretrained Language Models for Data-driven Financial Spam Filtering

박지혜, 조성준*
Summer Conference of Korea Data Mining Society (KDMS), Gangneung, Korea, 2023.

반도체 제조 도메인 특화 지식 그래프 구축을 위한 엔티티 추출 방법 연구 [View]

박지혜, 김현종, 박소형, 조성준*, 전성환
Spring Conference of Korean Institute of Industrial Engineers (KIIE), Jeju, Korea, 2023.

Twitter Embedding for Covariance Matrix Estimation & Portfolio Optimization

안재관, 박지혜, 유용재, 조성준*, 김규진
Summer Conference of Korea Data Mining Society (KDMS), Busan, Korea, 2022.

UMAP과 HDBSCAN 군집화 기반의 반도체 제조 공정 데이터 대상 레이블 전파 방법 [View]

박지혜, 최진우, 유용재, 조성준*, 정수호, 인창현, 박현섭, 조기윤, 윤동운, 김상혁, 조윤혁, 이은영, 이승호, 금의석
Spring Conference of Korean Institute of Industrial Engineers (KIIE), Jeju, Korea, 2022.

Automatic Keyword Extraction and NER-Normalization Based Knowledge Network Construction for Patent Summarization [View]

Hye Jin Lee, Sung Whan Jeon, Jihye Park, Sungzoon Cho*
Spring Conference of Korean Institute of Industrial Engineers (KIIE), Jeju, Korea, 2021.

A frequency- and ranking-based approach for new event detection using non-contiguous bigrams from text data of the central bankers’ speeches

Jihye Park, Hye Jin Lee, Sungzoon Cho*
Fall Conference of Korea Data Mining Society (KDMS), Seoul, Korea, 2019.

InfoAnoGAN: An Interpretable Anomaly Detection Method Using InfoGAN based on Unsupervised Learning

Jihye Park, Sungzoon Cho*
Spring Conference of Korea Data Mining Society (KDMS), Seoul, Korea, 2019.

Information Extraction, Data Labeling, and Dataset Shift Adaptation Algorithms for Big Data Analysis Model Construction

빅데이터 분석 모델 구축을 위한 정보 추출, 데이터 레이블링, 데이터셋 변화 적응 알고리즘 연구

National Research Foundation of Korea (한국연구재단)
September 2021 – August 2023
Scraped domain-specific textual data such as FOMC Statements and FOMC Minutes.

Research on building knowledge graph in the semiconductor domain

반도체 분야 지식 그래프 구축 연구

Samsung Advanced Institute of Technology (삼성전자 종합기술원)
February 2023 – June 2023

Research on deep learning based robust techniques for equipment anomaly detection

딥러닝을 활용한 강건한 설비 이상 탐지 기법 연구

Samsung Electronics (삼성전자 설비기술연구소)
December 2020 – June 2023

Text mining based hot topic detection from social media data

소셜미디어 텍스트 분석을 통한 투자 관련 핫토픽 탐지

NH Investment & Securities (NH투자증권)
September 2021 – December 2021

Development of automatic classification system for documents using artificial intelligence technology

인공지능 기술을 통한 문서 분류 자동화 시스템 구축

Samsung Electronics (삼성전자 혁신센터)
June 2019 – March 2021

Big Data Analysis

at Kyungbok University (경복대학교) on November 2022

Hot Topic Detection in Central Bankers' Speeches and Social Media

at College of Business Administration, Seoul National University (서울대학교 경영대학) on November 2021

Detection of word-changes in FOMC statements

at VAIV company (바이브컴퍼니) on August 2021

Hot Topic Detection in Central Bankers' Speeches for Text Mining-Based Early Warning Systems

at Korea Investment Corporation (한국투자공사) on October 2020

• I enjoy watching TV series about profiling such as Criminal Minds, Mind Hunter, and 시그널.
• I value a simple lifestyle and like to write down my thoughts and feelings without dwelling too much on the past.
• I am easily moved by warm words, and I genuinely like people. I believe that no one is purely evil. One of my favorite phrases is, "Love all, trust few, do wrong to none."
• I love music. My favorite artists include 윤하, Taylor Swift, and Chopin. I also enjoy playing the piano. Music always helps me chill out.