Naravich Chutisilp

Switzerland • 11/2024 - Present

• Designed, developed, and deployed a robust, real-time recommendation system for fixed-income derivatives using PyTorch, MLFlow, Optuna, and Airflow
• Achieved end-to-end latency of <1 minute to support rapid broker decision-making, generating >USD 655K in revenue
• Initiated and delivered a full-stack GitLab-integrated LLM bot for the CI/CD pipeline providing automated merge request summaries and code reviews
• The LLM bot enhanced development productivity across 50% of the data science team's projects

Switzerland • 09/2023 - 02/2024

• Built a production-ready Computer Vision system using satellite imagery for automated actuarial risk identification with JavaScript and Python
• Achieved >80% mIoU for critical solar panel segmentation, directly leading to securing an R&D budget >CHF 500k
• Engineered a highly efficient and scalable MLOps pipeline for satellite imagery segmentation using PyTorch, TIMM, and Lightning
• Enabled rapid processing of over 700 model architecture experiments from a single configuration, significantly accelerating R&D velocity

Switzerland • 10/2022 - 08/2023

• Developed a 6-DoF pose estimation and 3D object tracking system for carpentry tool heads
• Released an open-source C++ project using OpenCV and OpenGL. Manuscript accepted by MDPI's Applied Science Journal 2024
• Integrated the project into Augmented Reality (AR) software employing real-time SLAM
• Achieved high-precision performance with mean position and rotation errors of 3.9 mm and 1.19°, respectively

Thailand • 01/2021 - 04/2021

• Built an ML platform that accommodated financial data needs (>20 sources) and >5 data visualization requirements, using Azure, Spark and SQL
• The platform is now used by the marketing team to find leads for loans, contributing to THB 2.4MM of their revenue

Thailand • 06/2020 - 08/2020

• Implemented scroll virtualization of the product application, enabling >1,000 items to be rendered without lagging, using ReactJS and TypeScript
• Optimized page rendering by re-structuring state in Redux store and React state, increasing rendering time by 50%

Thailand • 06/2019 - 01/2020

• Built a dashboard for data visualization, helping Quantitative analysts to monitor their >1K alphas along with their >10 quantitative performance metrics using Python, Dash, and SQL
• Solved >20 GB of data syncing with multi-threading. Received a part-time offer after the internship
• Automated excel report generation for the Office Manager summarizing the performance of >10 employees and >10 specific key metrics, using Python, Pandas, and Numpy

Thailand • 01/2020 - 09/2021

• Developed interactive storytelling websites for digital marketing campaigns via different tech stacks including ReactJS, JavaScript, Svelte, Express, Flask, Firebase, Docker, Kubernetes, etc.
• One campaign attracted >1M organic users within the first day after launch and was #1 Twitter trending in Thailand, leading to the first 2 commercial contracts for the company
• Co-founded the company with a registered capital of THB 1M, which was earned from the commercial contracts

Thailand • 06/2018 - 08/2018

• Analyzed the unsupervised learning clustering algorithm to understand the meaning of each cluster with Scikit-Learn
• Compared similarity algorithms within the K-Means clustering project, assessing their impact on result variations