Danilo Onishi,巴西圣保罗州注册公司开发者
Danilo is available for hire
Hire Danilo

Danilo Onishi

Verified Expert  in Engineering

Artificial Intelligence (AI) Developer

Location
Registro - State of São Paulo, Brazil
Toptal Member Since
September 27, 2020

Danilo是一名机器学习工程师,拥有名古屋工业大学(Nagoya Institute of Technology)计算机科学硕士学位,在自动语音识别(ASR)领域有7年以上的经验。, natural language processing (NLP), and audio classification. 从研究到生产,他参与了机器学习生命周期的所有步骤. He also designs data pipelines and performs training, evaluation, 为教育和商业智能应用部署深度学习模型.

Availability

Part-time

Preferred Environment

Linux, Python, Vi

The most amazing...

...我建立的生产系统是一个语音识别后端,带有定制的声学和语言模型,服务200人,000+ requests per week.

Work Experience

Machine Learning Engineer

2021 - 2022
VoiceOps
  • 设计了基于bert的文本到文本模型的训练管道, including data loading, text preprocessing, training, validation, tracking of metrics, logs, and output files.
  • 培训和评估文本到文本模型,并根据公司使用的第三方服务和遗留系统对其进行基准测试.
  • 为遗留NLP系统编写新的后端api,将其迁移到由Hashicorp Nomad容器编排器管理的新基础设施.
  • 为基于nlp的语义搜索系统创建并实现了一个完整的数据管道, including the periodic check for new files to be processed, text embeddings generation and storing, and the back-end API.
Technologies: Machine Learning, Deep Learning, Generative Pre-trained Transformers (GPT), GPT, Natural Language Processing (NLP), Python, Back-end, APIs, Conda, Pandas, Databases, PostgreSQL, MySQL, Pytest, Nomad, Prometheus, Grafana

Machine Learning Expert

2020 - 2021
Online Learning Platform (Toptal Client)
  • 训练、评估和改进自动语音识别(ASR)模型. 与谷歌语音识别相比,最终模型的错误率降低了一半以上.
  • 设计了一个可扩展的后端API,可以接收任何格式和采样率的音频文件,并返回文本识别结果.
  • 将系统部署到GCP Cloud Run,并监控其结果. 截至2021年9月初,该系统每周有20多万次请求.
  • 用Python创建数据预处理工具,从GCP桶中检索客户端音频数据,并为ASR模型训练和评估做准备.
  • 使用TensorFlow为语音识别中的音频分类设计训练和测试工具.
Technologies: Python, APIs, Back-end, TensorFlow, Speech Recognition, Automatic Speech Recognition (ASR), Kaldi, Docker, Google Cloud Platform (GCP), Artificial Intelligence (AI), Machine Learning, Deep Learning, Linux, Bash, Git, Speech to Text, Pandas, Neural Networks

Software Engineer

2014 - 2019
Yume Technology Co. Ltd. (Assigned to Honda Research Institute Japan)
  • 共同撰写了一个系统,在2018年DCASE挑战赛的两个不同指标中获得了第三名和第六名的团队成绩, task 5.
  • 实现了一个服务器集群结构,与在单个服务器上训练模型相比,将语音模型训练速度提高了五倍.
  • 开始在公司使用深度学习模型进行语音识别, replacing legacy Hidden Markov models with DNNs, CNNs, attention models, and others.
  • 训练和评估英语和日语语音识别的声学和语言模型, targeting several applications such as robots, cars, and meetings.
  • 维护软件工具和框架的Git存储库,以便与其他开发人员共享,同时根据要求修复错误并添加新功能.
  • 分析和总结学术论文,掌握相关领域的最新技术,并评估哪些技术可以整合到团队使用的框架中.
  • 评估了一种音频去噪算法及其对语音识别性能的影响,并将其MATLAB实现移植到c++中.
Technologies: Generative Pre-trained Transformers (GPT), Natural Language Processing (NLP), GPT, Speech to Text, Data Science, Docker, Artificial Intelligence (AI), Python, Scikit-learn, NumPy, Kaldi, Chainer, TensorFlow, Git, Slurm Workload Manager, Speech Recognition, Deep Learning, Machine Learning, Bash, Neural Networks, Digital Signal Processing, DSP

Junior Development Engineer

2009 - 2010
Siemens Enterprise Communications (Now Unify)
  • 设计和组装微控制器编程工具,每件为公司节省2500欧元.
  • 评估客户支持团队报告的电话硬件问题,并提交固件修复程序.
  • 在国际发布之前,验证电话硬件的国家指定设置.
  • 在2010年中期作为工程实习生被提升到这个职位.
Technologies: PCB Design, Bash, Embedded C

声学事件识别系统的DCASE挑战2018,任务5

2018年DCASE挑战赛的第五项任务涉及训练系统识别日常活动的声学事件, such as cooking, speaking on the phone, or watching TV.

我们提出的解决方案是一个多任务CNN网络,在任务之间共享过滤器. 我负责修复和改进先前存在的TensorFlow实现, training and tuning the models, and writing the technical report.

对于已知的麦克风阵列,我们的实现获得了第三好的团队结果,对于未知的麦克风阵列,我们的实现获得了第六好的团队结果.

Environmental Sound Recognition Aiding Device

这个项目涉及一个环境声音识别设备,使用机器学习来支持听障人士. 它是在我攻读硕士学位时工作的实验室开发的. 我支持基于fpga的硬件重新设计(大约两张信用卡大小), 并最终为我的硕士论文制作了一个小型化的设计(SD卡大小).

Languages

Python, Bash, VHDL, C, C++, Embedded C

Tools

Git, Kaldi, Pytest, Grafana, Whisper

Platforms

Linux、Docker、亚马逊网络服务(AWS)、谷歌云平台(GCP)

Other

Machine Learning, Deep Learning, Artificial Intelligence (AI), Speech Recognition, Vi, Natural Language Processing (NLP), Speech to Text, Neural Networks, Automatic Speech Recognition (ASR), Conda, GPT, Generative Pre-trained Transformers (GPT), Digital Signal Processing, PCB Design, Embedded Hardware, Software Design, Slurm Workload Manager, FPGA, APIs, Back-end, Nomad, Prometheus, DSP

Libraries/APIs

NumPy, Scikit-learn, TensorFlow, Pandas

Frameworks

Chainer

Paradigms

Data Science

Industry Expertise

Telecommunications

Storage

Databases, PostgreSQL, MySQL

2011 - 2013

Master's Degree in Scientific and Engineering Simulation

Nagoya Institute of Technology - Nagoya, Japan

2005 - 2010

Bachelor's Degree in Electronics and Telecommunications

联邦科技大学-库里提巴,巴拉那州,巴西

Collaboration That Works

How to Work with Toptal

在数小时内,而不是数周或数月,我们的网络将为您直接匹配全球行业专家.

1

Share your needs

在与Toptal领域专家的电话中讨论您的需求并细化您的范围.
2

Choose your talent

在24小时内获得专业匹配人才的简短列表,以进行审查,面试和选择.
3

Start your risk-free talent trial

与你选择的人才一起工作,试用最多两周. Pay only if you decide to hire them.

Top talent is in high demand.

Start hiring