Projects

A collection of 20+ industrial and academic projects spanning search ranking, multimodal retrieval, GenAI workflows, speech/audio processing, EdTech ML, and creative AI — built across Bucketplace, RIIID, Humelo, and KAIST.

Figure 1: RCL contrastive learning modules from SURCL paper

Knowledge Tracing with Contrastive Learning

Proposed ACCL and RCL contrastive learning methods at RIIID, achieving state-of-the-art on student modeling across 6 benchmarks (dropout prediction, knowledge tracing). Deployed to Santa TOEIC platform.

ML Infrastructure: Model Registry & Pipeline (RIIID)

ML Model Registry, Dataset Pipeline & Infrastructure at RIIID (뤼이드)

Built ML model registry (MLFlow) and dataset pipelines (Airflow, Athena, BigQuery) at RIIID, serving 4+ products including SANTA TOEIC, IVYGlobal SAT, CASA GRANDE, and INICIE.

ML Pipeline Acceleration: GPU Utilization 25% to 95%

ML Pipeline Acceleration & Multi-GPU Training at RIIID (뤼이드)

Introduced RIIID’s first multi-GPU training, boosting GPU utilization from 25% to 95% and cutting initialization time from 1 hour to 10 seconds. Built CI/CD pipelines with GitHub Actions.

ICASSP 2019 Duration Controllable TTS: attention alignment and PDC model architecture

Emotional Text-to-Speech and Voice Conversion Systems

Led development of duration-controllable TTS and emotional voice conversion at Humelo, producing two ICASSP publications (2019 Oral 1st author, 2020). Won Minister of Science and ICT Special Award at K-Startup 2018.

Government & Public R&D Grant Management

Secured ~US$875K across three competitive Korean government R&D grants (IITP, TIPS, Seoul R&BD) at Humelo, covering brain-inspired AI, emotional TTS, and voice conversion research.

CBRNN architecture with transfer learning for polyphonic sound event detection (ICASSP 2019)

Polyphonic Sound Event Detection with Transfer Learning

Developed convolutional bidirectional LSTM with synthetic data-based transfer learning for polyphonic sound event detection at Humelo, achieving +28.4% F1 improvement. Published at ICASSP 2019 as corresponding author.

AI Music Composition and SM Entertainment Collaboration

Led AI music composition and rap synthesis at Humelo, presented at SXSW 2019, collaborated with SM Entertainment and rapper Sleepy (KBS Documentary), and received coverage from 10+ national media outlets.

Speech Emotion Recognition & Classification System

Built a multi-class speech emotion recognition system at Humelo using SpeechCNN and CRNN architectures with MFCC/Mel-spectrogram features, integrated into the Emotional TTS pipeline.

Overall system architecture for EEG signal classification with Deep RL (IEEE SMC 2018)

Attentional Control for Time-Series Data (Master's Thesis)

Master’s thesis at KAIST on attentional control for time-series classification and synthesis, solving the memory-based vs. memoryless trade-off for EEG signals. Oral presentation at IEEE SMC 2018.

Hunter-prey model and Deep Boltzmann Machine diagrams for multi-agent cognitive policy learning (KIIS 2016)

Multi-Agent Cognitive Policy Learning through Competition

Undergraduate research at KAIST on multi-agent cognitive policy learning through competitive reinforcement learning, demonstrating emergence of complex behaviors. Won Best Paper Award at 2016 KIIS Conference.