Career Profile

Machine Learning Engineer, yearning for solving interesting problems with machine learning technologies. Interested in Deep learning, Natural language processing, and Generative models.

Experiences

Machine Learning Engineer

2019.1 - Present
Hyperconnect, Seoul

Worked on various machine learning related tasks, including but not limited to:

  • Open-domain Conversation Chatbot
  • Recommender system
  • Few-shot face reenactment
  • Text classification
  • SIMD/Assembly-level optimization of low-bit neural networks

Machine Learning Team Tech Lead & Machine Learning Engineer

2017.3 - 2018.12
Scatter Lab, Seoul

Worked as a tech lead for machine learning team, developing machine learning models for open-domain conversation solution Pingpong. Developed several models including but not limited to:

  • Retrieval-based chatbot
  • Generic response generation via Text/Conversation classification
  • Semantic text similarity
  • Text generation

Undergraduate Research Intern

2016.12 - 2017.3
Vision & Learning Lab, Seoul National University

Worked on pedestrian detection task, supervised by Gunhee Kim.

Selected Publications

Beomsu Kim*, Seokjun Seo*, Seungju Han*, Enkhbayar Enkhee*, Buru Chang.
EMNLP 2021 (Findings)
* denotes for equal contribution.
Seungju Han*, Beomsu Kim*, Seokjun Seo*, Enkhbayar Enkhee*, Buru Chang.
Arxiv 2021
* denotes for equal contribution.
Embedding Normalization: Significance Preserving Feature Normalization for Click-Through Rate Prediction
Joonyoung Yi, Beomsu Kim, Buru Chang.
ICDM NeuRec 2021, Best Paper Award
Sungjoo Ha*, Martin Kersner*, Beomsu Kim*, Seokjun Seo*, Dongyoung Kim.
AAAI 2020
* denotes for equal contribution in alphabetical order.
Junhyug Noh, Soochan Lee, Beomsu Kim, Gunhee Kim.
CVPR 2018

Projects

Entity Resolution for Tweets (2016) - Suggested model for Entity Resolution task on tweets, done by undergraduate research opportunity program on SNU IDS Lab.
Timetable Generation Problem (2015) - Post-enrollment & Student sectioning timetable generation problem for Gyeonggi Science High School.