Summary
Overview
Work History
Education
Skills
Awards and Scholarship
Conferences Presentations
Languages
Timeline
Generic

Tomoaki Hayakawa

Nerima

Summary

Image Recognition Engineer with over 3 years of experience in computer vision and deep learning. Proficient in research and development of deep learning models for object detection, image classification, and anomaly detection. Successfully deployed image recognition models for various industries to on-premises environments, including edge devices requiring real-time processing, and implemented them in practical field operations.

Overview

3
3
years of professional experience

Work History

Image Recognition Engineer

NEC Solution Innovators, Ltd. (NEC Group)
2021.04 - Current
  • Led a cross-functional team of image recognition engineers, frontend developers, and backend engineers to successfully launch an application for detecting and recognizing food items in images within the set deadline
  • Reduced evaluation time by 30% by developing assessment tools for image recognition models
  • Doubled inference speed of image recognition models through optimization of pre-trained models, enabling real-time processing in on-premises environments
  • Improved object detection model accuracy by 10% through implementation of novel algorithms and expanded training data using semi-automated annotation, resulting in broader use cases and acquisition of new clients
  • Validated CI/CD implementation to streamline development of existing products incorporating image recognition models, significantly reducing testing time compared to conventional methods
  • Contributed to company's AI/Computer Vision initiatives by sharing internal reports on latest technologies and services in the field
  • Designed and developed an image recognition model for authenticating pharmaceutical images, achieving 80% accuracy with a limited dataset of only 10,000 images

Education

Master of Science - Computer Science, Speech Recognition

University of Yamanashi
Yamanashi, Japan
03.2021

Bachelor of Science - Computer Science, Speech Recognition

University of Yamanashi
Yamanashi, Japan
03.2019

Skills

  • Python, C/C++, Perl, MATLAB
  • Git, Github, Docker
  • Bash, Zsh
  • Tensorflow, Keras, Pytorch, Chainer, ONNX

Awards and Scholarship

  • Awarded Best Presentation for Master's Theses(2021)

Conferences Presentations

  • Tomoaki Hayakawa, Chee Siang Leow, Akio Kobayashi, Takehito Utsuro, and Hiromitsu Nishizaki, “Language and Speaker-Independent Feature Transformation for End-to-End Multilingual Speech Recognition,” Proceedings of INTERSPEECH2021, pp.2431-2435, 2021.8 https://www.isca-archive.org/interspeech_2021/hayakawa21_interspeech.html
  • Chee Siang Leow, Tomoaki Hayakawa, Hiromitsu Nishizaki, and Norihde Kitaoka, “Development of a Low-Latency and Real-Time Automatic Speech Recognition System,” Proceedings of the 2020 IEEE 9th Global Conference on Consumer Electronics (GCCE 2020), pp.464-467, 2020.10 https://ieeexplore.ieee.org/document/9291818

Languages

Japanese
Native language
English
Beginner
A1

Timeline

Image Recognition Engineer

NEC Solution Innovators, Ltd. (NEC Group)
2021.04 - Current

Master of Science - Computer Science, Speech Recognition

University of Yamanashi

Bachelor of Science - Computer Science, Speech Recognition

University of Yamanashi
Tomoaki Hayakawa