곽기은

스킬: 데이터 분석, python, 머신 러닝, 커뮤니케이션

< Job description > ❏ 데이터 분석 프로젝트 실무 ❏ 자회사들의 pain point 발굴 및 In-house 컨설팅을 통한 DT 혁신 ❏ 그룹사의 digital roadmap 설정 < Projects > ❏ LLM을 활용한 Chemical 공장 안전 관리 강화 - 산업안전관리의 일환, 작업 전 필수인 Job Security Analysis 보고서 작성 추천 시스템 개발 - Dynamic few-shot selection, RAG, semantic element extraction ❏ 초분광 카메라를 통한 이물 검증 - 초분광 카메라의 약 200가지 분광을 바탕으로 사물을 판단하고, 사물 뒤에 숨겨진 물질까지 파악하는 AI 모델 개발 ❏ B2C 부동산 어플 개발 참여 - 학군, 교통, 지역, 호재, 주변 시설, 아파트 정보, 호가 등을 활용하여 공동 주거지역의 적정 가격 측정 모델 개발 (부동산 AVM) - 부동산 관련 feature mart 구축을 통해 1000개 이상의 feature 표준화

quipped with advanced AI skills by being the middleman between the Singapore regional center and the local Seoul branch. Worked with local insurance experts to develop data analytics service to Korean primary insurers. < Projects > ❏ Insurance claim fraud detection (Motor insurance) - Developed automated process to read / process / standardize data, and provide fraud probability score based on machine learning and fraud rules ❏ Parametric product based on rainfall return period - Developed a Value At Risk (VAR) financial model to calculate return period for rainfall for each weather station, in order to quantify flood risk for each city in Japan ❏ Motor Insurance Automatic Quotation System Using Artificial Intelligence - An end to end vehicle claim processing app using image recognition and semantic segmentation. Took part in developing the AI model that identifies damages from images taken by the end users. - https://www.munichre.com/en/solutions/reinsurance-property-casualty/vahana-ai.html ❏ Commercial Fire Insurance Target Marketing through Customer-Specific Scoring - Utilized building data, financial data, as well as natural risk scores that covers 8 million business operators and 7 million buildings in Korea, to analyze the factors that influence fire risk of each individual buildings in Korea ❏ Predicting loss ratio of potential individuals (motor insurance) - Collaborated with one of the telematics digital insurance player in Korea to develop statistical model that would predict the loss ratio of potential customers before going into the quotation process

학점: 3.8 / 4.0

학점: 3.7 / 4.5