디큐

스킬: typescript, python, React, SQL, Google Cloud Platform, 커뮤니케이션, 데이터 분석, 리더십, 문제 정의하기, 머신 러닝, 딥러닝

Built an end-to-end data ingestion pipeline from scratch, integrating Google BigQuery and Google Cloud Run Designed an effective User Behavioral Log Schema and implemented a log proxy server utilizing Google Pub/Sub and NestJS. Constructed a robust DBT pipeline, establishing a data mart layer for enhanced analytics. Orchestrated the development of a business insight tool using Metabase. Cultivate data driven culture by publishing data magazine bi-weekly basis, automate KPI monitoring slack bot Pioneered the creation of a Generative AI-adapted pipeline for producing Generative Interaction Videos from text inputs. Leveraging LLM, reasoning material and texture based on image/video and generate haptic response. Implemented a real-time human emotion analyzer and analyzing dashboard, integrating Google mediapipe.

Designed and deployed a recommendation engine, enabling personalized content feeds through Deep Learning techniques (lightGCN) Resolved cold start issues with innovative solutions including ALS method application Pioneered the development of the entire backend system utilizing NestJS and GraphQL Established a highly automated CI/CD pipeline leveraging GitLab Runner Initiated a Daily KPI report Slack bot for tracking vital metrics such as revenue, GMV, DAU, and conversion rates Fostering an agile development culture, led Scrum methodologies, and managed sprint retrospectives Set up BI tools (Redash, Metabase) for comprehensive data analysis and integrated third-party MarTech tools into a React Native app

- Developed in-house distributed data processing and scheduling system (Spark on Kubernetes) which processes over 50m rows of online class transaction data on daily basis - Led and established data team - Reduced ad-hoc query requests on data analysts by 40% - Published internal article about data analysis weekly - Enhanced reliability of Spark batch processing jobs by 50% - Implemented recommendation system - GNN, RF, RNN based - tensorflow, scikit-learn - Batch Data preprocessing for Recommendation System - AWS cloudwatch, AWS Batch - Built ETL pipeline ( from MongoDB to AWS Redshift ) - pySpark, AWS EMR, Airflow - Built Event Streaming pipeline ( AWS Kinesis ) - AWS Kinesis Firehose

Designed API server / DB Schema for the enterprise collaboration tool Built a developer team culture based on agile Automated the CI/CD pipeline using Gitlab-runner, Docker

Developed a Sarcopenia diagnosis module using the ratio of CT numbers of bone, Subcutaneous fat, and Visceral fat (Avg. inference speed: 0.03sec/image, 95% Accuracy on test set) Developed a Scoliosis diagnosis module Designed a database for medical data and automated conversion pipeline for encrypting sensitive information Modified the open-source medical viewer(horos) to make annotation and data gathering easy Constructed a server, network system for deep learning training Developed a MPR(Multi-Planar Reconstruction) software for CT images

Developed a Emphysema of the lungs diagnosis module based on X-ray images (99% accuracy on test set) Developed a Pressure Fracture of the spine diagnosis module (90% accuracy on test set) Developed a Periodontal Disease (Cyst, Periodontitis) diagnosis module (Cyst module: 99% accuracy on test set) Developed a Ossification of Posterior Longitudinal Ligament(OPLL) diagnosis module Developed a annotation helper software for medical images Optimized a hospital network for quicken the data I/O speed Automated a data gathering from the doctor’s computer (Picture archiving communication system(PACS))

Built a cross-platform application using ionic framework (Angular based) Managed the development team resource and designed technical stack in general manner Developed a chatbot system using Kakao Open API

SNU BIBS Lab. Refined a dataset from NCBI for finding bio marker of 18-class cancer Automated a data crawling process (web crawling library) Visualized the relation between clustered gene data so that researchers can easily pick the data to be analyzed deeply SNU AIEES Lab Collected a dataset of energy consumption data of SNU buildings Researched the relation between average temperature and greenhouse gas emission in SNU

Full stack recommendation system coding guide FastAPI - python React - javascript

파이썬, 리액트, 깃허브 액션에서 스타트업 개발자 성장 노하우까지