Upon Toyota group companies' shift to software 2.0, I launched and led a cross-team project of AI SDK and AI pipeline for contributing to expanding the Toyota group's projects (e.g. AD/ADAS, Woven City, etc) that utilizes ML/AI and helped their success by effectively utilizing AI/ML for their domains with the following efforts:- Providing python AI SDK that can be utilized as SW asset pools and cumulative knowledge pools for helping general ML developers and perception engineers' productivity in Woven, Toyota and Denso without re-inventing-the-wheel per each project.- Providing end-to-end ML pipeline and MLOps solution with ML workflow engines such as Kubeflow, which enables continuous improvement of ML models and datasets.- Working with AD/ADAS development team, the product and service are heavily used by product development of AD/ADAS applications.- Launching CI/CD pipelines with GitHub Actions for testing, integrating and delivering the ML/AI-related services and products.

As AI/ML becomes crucial for AD/ADAS development, the importance of the collection of large amounts of sensor data for training and testing AI/ML models has been increased. In this role, with my background as an AI developer, I have contributed to developing data processing and storage SW module in vehicle with C++ for effectively collecting sensor data from vehicles to servers so that the performance of AI/ML models can be continuously improved with closed-loop AI/ML development workflow. I published 4 granted US patents as 1st inventor, and with the acknowledgement of the importance of the project, one of my patents was awarded with CTO Award 2021 (Koibuchi Award) as 1st inventor.

- Continued role from Toyota Motor Corporation. Please refer to its section for details.- As another project, I also launched and led a new project team from inception through production. I developed a new vehicle application for safely and comfortably avoiding vehicle collision with turn signal recognition with Neural Networks. This project started from my personal research first, and after PoC, I launched and drove a product development team (up to 5 people). The new feature was successfully deployed to market vehicles on July.28.2021.- I architected, designed and implemented deep learning algorithms, NN training programs, data curation and active learning programs using Python and PyTorch.- I architected, designed and implemented vehicle software with neural networks complying to vehicle safety standards such as AUTOSAR and MISRA, using C++, TensorRT, NVIDIA CUDA and other NVIDIA tools and QNX for automobile embedded systems with vehicle ECU and NVIDIA chip.- I designed and launched an application-specific ML-Ops pipeline that is connected to Vehicle ECU using Jenkins, Groovy, and AWS.- I published 7 granted US patent as 1st inventor- Four prizes awarded for the patents of this project: - "Invention of the Year 2020" (as co-inventor) - "Invention of the Year 2019" (as 1st inventor) - "COO Special Award (Mushigami Award) (2019)" (as 1st inventor) - “2nd Most Invention Proposals Award (2019)“ (as 1st inventor)

Please refer to the description about the role "Staff Engineer of Automated Driving".

My role was to launch in-house AI/ML applications for an AD/ADAS product for the first time in Toyota. I launched a software/perception AI development team (up to 10 people) which was responsible for developing Object Detection, Road Recognition and AI related features for AD/ADAS with production-level quality using C++, Python and Pytorch, and I served as a team leader of the team. The launch effort includes:- PoC software development with NVIDIA Drive PX2 (DDPX)- Team building with Agile Scrum.- Launching software development process with dev tools.- Involving other teams in Toyota (e.g. downstream application development teams) for rapidly integrating our software components into the existing vehicle system.After the team grew, the team was splitted into three teams, and I led one of the teams: Object Recognition team. After the PoC phase, the team shifted to product development complying to vehicle safety standards such as AUTOSAR and MISRA, and its product was finally and successfully shipped in 2021.- I continued the same responsibilities after moving to the new position "Toyota Research Institute - Advanced Development, Inc."- Published 8 granted US patent as 1st inventor

- Responsible for technology survey, product planning, and advanced development at CTO office.- Led cross-functional team to launch novel products involving US, Japan, Israel, and UK.- Developed server software Proof of Concept (PoC) working with datacenter customers by using C++, C, and Python.- Researched future technology trends and proposed technology roadmap to company headquarters.- Published 33 granted US patents as 1st inventor and 1 paper as co-author.

- Designed embedded Solid-State Drive (SSD) software (developed via C).- Designed SSD utility software (developed via C++) for Windows and Linux.- Implemented test codes and product line code (developed via Java and Lua).- Led cross-country collaborations with US controller vendors.- Designed manufacturing software and launched manufacturing production lines in China, the Philippines, and Taiwan.- Published 31 granted US patents as 1st inventor.

- Designed Persistent Semiconductor Memory circuit (Ferroelectric Random-Access Memory, FeRAM) at transistor and logic gate level.- Designed automated testing and analytics software to evaluate FeRAM using C++, shell, and Perl. Software automatically detects failure circuits and processes by analyzing big data from FeRAM chips’ test log.- Published 5 granted US patents as 1st inventor and 3 papers as co-authors.