Experience

• 

  Huawei

  Jan 2011 - Jul 2011

  Electrical engineer
• 

  Ntt research

  Apr 2015 - Mar 2017

  Data scientist

  Participate in the Project #Todai Robot Project means "Can a Robot Get Into the University of Tokyo?"- Compared the GCP speech recognition and Kaldi(DNN model) accuracy- Built the ensemble Deep learning model(LSTM, Word2Vec) for the English Listening Conversation Question- Had a presentation scheduled in every month to collaborate with professors from another university.

• 

  Fujitsu

  Mar 2019 - Aug 2020

  ①：Developed the clothes top-bottom recommendation (LSTM) system with the flask and react based on AWS - Create the supervised image dataset labels - Optimize parallel data processing - Utilize the fashion-related ML papers (TensorFlow) to implement our need for the model - Tech stack: React(js), Flask(python), Tensorflow 1.0, AWS(Serverless Lambda, API Gateway, DynamoDB), docker-compose➁：As a data scientist to process/analyze structured data and implement models using DL(PyTorch) and ML (collaborative filtering) proposals in course recommendation projects. - Tech: python, pandas, PyTorch Show less

  • Software Engineer, AI platform Dept.

    Jun 2019 - Aug 2020

  • ML Engineer, Incubation Center Dept.

    Apr 2017 - Jun 2019

  • ML Engineer, Incubation Center Dept.

    Mar 2019 - Mar 2019
• 

  Ridgelinez株式会社

  Aug 2020 - Dec 2023

  Consulting senior software engineer, data engineer, mlops engineer. tech group

  Apply the End-to-End development service to help the customers to adapt to the agile methodology and gain business value quickly.① Report to VP .As a tech leader designed and built the Finance Validation Process and Data-driven(ML, NLP, OpenAI GPT3). Data lake,data quality(Great expectation),ETL, and data warehouse which integrated with API and Tableau. Also, include the security, CI/CD pipeline and infra for the global manufacturing company.(Lambda, Glue, s3, Aurora)②：Report to C-suite Executive.Designed and built the automated workflow (data processing, ML pipeline, env checking) integrated with Palantir Data platform using Airflow based on AWS CI/CD in the large global FP&A project - Tech stack: python, shell, airflow(ECSoperator), mlflow, AWS(EC2, ECS, EFS, S3, ECR, Codecommit, CodeBuild, CodePipeline, Lambda, NW, Sagemaker) ③: Working as a software engineer to handle the backend APIs service development with customers by agile methodoloy on Azure devops platform. - Datahub APIs integrated with SAP S/4HANA, Cydas, and Outsystem - Tech stack: python, SQL, Azure(serverless function, cosmos DB, SQL, AzureDevops pipeline, Kanban) Show less

• 

  Sambanova systems

  Jan 2024 - now

  Principal customer engineer

  Unicorn Startup@Palo AltoFor Japan, Singapore and Taiwan client, demonstrate, architect, design, and deliver solutions with GenAI SambaNova Systems products with appropriate value propositions and technical merit- Conduct LLM SFT/RHLF/Contious pre-training for the customer to ensure alignment with their specific business requirements.- Conduct the fine-tuning of LLava for the customer and advise on the RAG system design with the customer.- Three-month PoC for the customer to implement domain-specific knowledge fine-tuning and create the RAG system to adjust the QA bot. - The first technical win in Japan Region LLMs/Llava/DL/ML, NLP/CV/Object Detection, MLOps, E2E development. Show less