Sung-ju Kim

Machine Learning Scientist, Engineer

github linkedin

Profile


Career

2023.1-Current AI Researcher at Naver Cloud.

2018.7-2022.12 AI Researcher at Naver.

2016.2-2018.7 AI Researcher at Nomad Connection.


Education

2014.3-2016.2 M.S. in Computer Science & Engineering at Kyungpook National University

2009.3-2014.2 B.S. in Biotechnology at Kyungpook National University


Interests

Machine Learning, Computer Vision, Natural Language Processing, Startup


Publication

S.-J. Kim and S.-Y. Park ”Lane-Level Positioning based on 3D Tracking Path of Traffic Signs,” in Visapp, 2015. (International Conference)

S.-J. Kim and S.-Y. Park ”Lane-Level Positioning based on 3D Tracking Path of Traffic c Signs,” in CICS, 2015. (Domestic Conference)

S.-J. Kim and S.-Y. Park ”Traffic Sign Recognition Using SVM with Generated Virtual Images,” in IPIU, 2015. (Domestic Conference)


Projects

Clova AI App/Speaker, at Naver in 2018~Current

Developing dialog management engine of Clova AI App and Clova AI Speaker.


Intent Classifier and Sentence Encoder for Chatbot (in Shinhan bank app named SOL), at Nomad Connection in 2018

Intent Classifier for classifying queries and Sentence Encoder for matching queries with similar pre-defined questions. Sentence Encoder is used to route ambiguous user queries to pre-defined questions which chatbot can solve.


ID Card OCR Recognition for Non-face-to-face Authentication, at Nomad Connection in 2017

Korean, English, Special Character Recognition with Convolutional Neural Network using character hierarchy. Identification card detection using restricted hough space. Character line detection with several computer vision methods.


Chatbot Solution, at Nomad Connection in 2016

wit.ai like chatbot build and service system. Several NLP and Machine Learning Techniques are used for Named Entity Recognition, Query Classification.


Sentiments Analysis for Financial News, with QUARTERBACK INVESTMENT at Nomad Connection in 2016

Sentiments Analysis Engine for Financial News data using Convolutional Neural Network.

this sheet shows a part of analysed news data


Keyword Extraction Algorithm with Financial Analytic Perspective, at Nomad Connection at 2016

Extract meaningful keywords from news, analytic reports and disclosures using Word2Vec, TF-IDF, Support Vector Regression(SVR), General NLP Technique. Word2Vec & SVR is used to express perspective of analyst.


Smart Oil Filter Sensor, Server and App, at Qualc in 2015~2016

Smart Oil Filter Sensor, Server and App. Made with Arduino Open Source Hardware, Hybrid App using phonegap, Flask-RESTful for server


Tracking of 3D Path of Traffic Sign, at KNU in 2014~2015

Detecting Traffic Sign and Finding 3D path of Traffic Sign using Stereo Camera. Support Vector Machine, Multi-Layer Neural Network, Stero Vision and Several Computer Vision Techniques ared used


Bin Picking System, with ETRI at KNU in 2015

Bin Picking System for Robot. I took the bin detection part. Multi Layer Neural Network is used.


Traffic Sign Detection & Recognition, with ETRI at KNU in 2014

Traffic Sign Detection & Recognition using Support Vector Machine, Histogram of Oriented Gradient Feature and General Computer Vision Technique.

...


Prizes

...