<h1>

Seita Fujiwara.

I am Machine Learnig Engineer |

</h1>

About Me

I have professional experience in Computer Vision and Time Series Analysis within machine learning. In my free time, I explore LLMs, participate in Kaggle competitions, and continuously expand my knowledge. Recently, I have developed a strong interest in AWS and have been actively studying cloud technologies.

Experience

connectome.design

(ML Engineer)

From 2024-02 to Current

Defect Detection in Products:

Developed a defect detection system for consumer products using YOLO and MMDetection, including dataset preparation through manual video annotation. Integrated a custom tracking algorithm to enhance detection accuracy and adapted the system for real-time performance by optimizing code and logic.

MMDetect Machine Learning YOLO
Prediction and Evaluation of Power Equipment Failures:

Developed machine learning pipelines to identify failure causes in power equipment using signal processing techniques and time-series models such as LSTM. Refactored legacy codebases to improve model accuracy and functionality. Deployed AI models on AWS and integrated them with the frontend using serverless architecture.

Docker Machine Learning AWS LSTM Lambda

R Systems

(Support Engineer)

From 2023-06 to 2024-09

Support Engineer:

Primarily used tools such as Kibana and Datadog to monitor and analyze errors in Mobile Payments. Communicated with companies and telecom operators during incidents, escalating issues as necessary. Additionally, handled customer user registrations and other tasks using Linux

datadog kibana