On-Device Speech Recognition- Model Compression- Knowledge Distillation (KD)RNN-T based on-device ASR- model training, optimization, quantization (pytorch, TFLite)- android application- online decoder LAS-CTC based E2E ASR- model training, optimization- decoderOpen source- Pytorch Korean Tutorial TranslationDeep learning based text classification- modeling and inference speed optimization - tensorflow, pytorch, c#Language Model for Automatic Speech Recognition Service- Text data crawler - Text tokenizer considering pronunciation loss- RNN based language model- ASR performance optimization for the specific service domain- Service Domains => POI (place of interest) search for navigation service => Music search => Schedule, weather, etc. Main projects- ASR speaker (NUGU)- Speech assistant for car navigation (Tmap) - SMS spamfiltering- RNN-T based on-device E2E ASR for car navigation- T-phone Call Record Speech Recognition using E2E- E2E ASR Knowledge Distillation- IOS Phone Recoding ASR