Algomatic is a company that develops and provides services utilizing generative AI technologies, such as large language models.
I am working as AI engineer.
I work an AI/LLM engineer and Data Scientist.
AVILEN is a Japanese AI and machine learning company that provides AI-driven solutions, consulting services, and educational programs. AVILEN is a publicly listed company.
I'm one of DS-Hub members. https://avilen.jp/dshub/
The AVILEN DS-Hub is a community of machine learning researchers who are actively implementing AI technology in society through AVILEN projects. Comprising 200 members, these individuals have passed a rigorous independent test with a 6% pass rate, prepared by AVILEN.
https://company.baseconnect.in/
I am working part-time as an AI software engineer developing our in-house service to improve the accuracy of LLM. This job particularly requires knowledge of natural language processing. I am involved in optimizing LLM prompts, managing the accuracy and quality of outputs, and also working in LLMOps.
What I emphasized the most was optimizing the AI model to meet business needs. I maintained close communication with the marketing team to minimize development and operational costs while maintaining a sufficient level of performance. In business, ROI (Return on Investment) is critical, even when academic value is present. Therefore, I carefully balanced the performance of the model with the costs of development and operation to ensure that the product remained feasible. Specifically, I selected lighter models to reduce operational costs without sacrificing accuracy, achieving a 2 to 4 times cost reduction. Furthermore, I held regular meetings with the CEO and marketing team to facilitate decision-making, ensuring the project stayed aligned with business goals.
Keywords: Python, LangChain, Dify, Langfuse, WandB, Docker, AWS, Snowflake, Pytest, Fine-tuning LLMs, Docker.
https://www.seetruetechnologies.com/
Worked on an object detection project using video data obtained from devices produced by the company. The YOLO model was trained on a custom dataset and I developed a simple web app using Python that allows clients to easily start the object detection process and analyze the results. The developed product was demonstrated for clients.
The challenges and achievements during this internship involved integrating the AI model I trained into a prototype web application, allowing anyone to execute the model with any dataset. This enabled us to showcase the product’s value to potential customers. The difficult part was aligning the nature of the training data with the actual inference data and the product environment. By communicating and matching the data from the actual environment, I focused on maximizing the AI model’s value.
Keywords: YOLO, Python, Streamlit, Dash.
I developed a web application aimed at improving internal workflow efficiency. Python was used for web development, and basic machine learning methods such as random forests were integrated to enable seamless AI integration accessible even to non-technical users. SHAP was utilized to visualize feature contributions, and MLflow was employed to manage the experiments.
I believe that a data scientist's role is not merely to analyze data and build machine learning models but also to identify business challenges from a business perspective and propose optimal data analyses and models to address them. During my internship at SystemCreate from January 2022 for three months, I engaged in various tasks as a data scientist, and I felt a great sense of fulfillment contributing to the business through data analysis. It was incredibly rewarding to see how the insights I derived from the data led to business growth. This experience made me strongly recognize the significance of data scientists. Moving forward, I aim to continue tackling similar challenges and leveraging data to deliver value to businesses.
Keywords: Python, Scikit-Learn, Dash, Web-application, Random-forest, Figma, Canva, GitLab.
We formed a team of seven students and participated as an engineer in an activity that was primarily a mobile app as a business. I was mainly responsible for the server-side system. We eventually won the top prize for this activity in the Business Plan Contest held in 2021. I have experience in developing services, mainly mobile apps. I was mainly responsible for the server-side system: we built a push notification system using AWS.
Keywords: Firebase, AWS, Amplify, GitHub, Python, Dart, Flutter, Figma.
Automatically extract ML experimental conditions from your Python scripts with GPT4, and save them via WandB.
Ever found yourself lost in a maze of changing experiment conditions in your early ML scripts? 😵 No worries—here’s the solution you’ve been looking for! ✨
Project page: https://logllm.tiiny.site/
I am compiling papers on LLM in tabular format for my own research. The purpose of this repository is to particularly organize papers from various fields in a comprehensive table. The repo gets over 170 stars.
https://github.com/shure-dev/Awesome-LLM-related-Papers-Comprehensive-Topics
Our team developed a web application as part of a university course abroad. We aimed to create software that allows users to interactively handle data. In the future, we planned to add features such as principal component analysis. Using NoSQL and Dash, we focused on building a lightweight, interactive application. We developed a dashboard to make data analysis easy.
I participated in the Numerai financial time series forecasting competition, which involves predicting stock prices. I generated my own features and used classification models to make predictions.
https://numer.ai/~shure, https://signals.numer.ai/~shure
What is Numerai? > https://numer.ai/home