Data Scientist (AI-Ready / Tacit Knowledge Analysis)Translated from Japanese

Data Scientist (AI-Ready / Tacit Knowledge Analysis) Recruitment Overview ■ Main Responsibilities Facing a severe labor shortage due to a declining birthrate and aging population, Japanese manufacturing industries are experiencing a structural risk where tacit knowledge (intuitive judgments based on experience, fine motor control, and force adjustments) acquired by skilled workers over many years is being lost upon their retirement. Simultaneously, as research and development in Physical AI accelerates, particularly in the US and China, the importance of digitizing this on-site tacit knowledge and converting it into a form that AI can understand and inherit is rapidly increasing across the industry. Previous efforts in related fields have been limited to preparing data from production systems and applying OCR/RAG to existing unstructured data such as blueprints and manuals. These are all preparations for utilizing "data that already exists in some form." For "tacit knowledge that does not exist as data in the first place," such as worker movements, eye movements, and reasoning processes, methods for collection and structuring have not yet been established. To address this unexplored area, a new methodology is needed that collects multimodal data, including linguistic data based on interviews, and video, keypoint, and eye-tracking data, and then identifies and structures tacit knowledge not present in manuals through differential analysis between skilled and general workers. As a professional in AI data, FastLabel has been selected for the national project "Research and Development of AI-Ready Data for Manufacturing Industries (GENIAC)" by the Ministry of Economy, Trade and Industry (METI) and the New Energy and Industrial Technology Development Organization (NEDO). We will establish the technical foundation and methodology in this field as data infrastructure to support the knowledge of skilled workers, which is not verbalized on-site. ​ In this position, you will be responsible for the entire process as a Data Scientist, from designing data collection methods for production sites and data analysis to extracting tacit knowledge and verifying technologies for automating analysis and extraction using AI models such as VLMs. You will lead the analytical technology for the new domain of tacit knowledge and contribute to the launch and expansion of a business that supports skill inheritance and AI-readiness in a wide range of industries, starting with manufacturing. 《Specific Responsibilities》 Upon joining, you will be responsible for the research and development project led by our company to make tacit knowledge from manufacturing sites AI-ready. The core mission is to collect the words (tacit knowledge) of skilled workers as primary information, design data collection methods to capture it, link it to features in data such as work videos, keypoints, and eye movements through analysis, and translate it into structured data that AI can learn. 【Our Research and Development Project】 ・Interviews and Verbalization of On-site Tacit Knowledge: Visit factories with PMs and consultants to extract and verbalize knowledge not found in manuals through interviews with skilled workers. ・Data Collection Design: Based on interviews, define the target processes, tacit knowledge to be extracted, sensor configurations, and evaluation criteria. ・Extraction of Tacit Knowledge: Extract features where tacit knowledge appears by comparing work data of skilled and general workers. ・Automation Model Development and Evaluation: Design hypotheses for automating the tacit knowledge extraction process and verify hypotheses through prototype development. *This includes tasks such as fine-tuning VLMs based on manually analyzed results. *In addition to this project, you may be assigned to internal development projects in recognition AI and robotics AI, or development projects commissioned by clients. Scope of Change: After gaining experience in the above duties, your responsibilities may change to encompass all company operations depending on your aptitude and preferences. ■ Background for Recruitment To strengthen our personnel structure due to business expansion. We are seeking a technical professional to lead AI data projects in the highly challenging domain of tacit knowledge, enabling us to lead more clients to business success. As AI technology development and utilization advance across all industries, you will develop foundational technologies in new data domains that can lead to AI robotics. *Coding challenges may be administered as needed. ■ Appeal of the Job 【Deep Understanding and Practical Skill Acquisition in Data-centric AI】 While the importance of data in AI model performance is increasingly recognized, opportunities to meticulously handle data from data collection methods and accuracy, and to practice through to VLM model training and evaluation are rare. This project requires addressing data quality from various angles, and the insights gained in Data-centric AI are expected to be extremely valuable and highly competitive in the labor market. 【Experience Applying Cutting-Edge Technology to the "Real World"】 Beyond simply tackling technically challenging themes with novelty, we emphasize that these challenges are based on the needs and high interest of major manufacturing companies that support Japan. You can participate in initiatives aimed at contributing to the competitiveness of Japanese manufacturing, with the full cooperation of world-class production sites. 【Application to Robotics AI】 FastLabel will launch a Robotics AI Business Unit in 2026 and focus on robotics AI. Currently, AI control for robots is limited to aiming for the average human work quality, but it is highly realistic that a trend will emerge in the future to replicate the work of skilled individuals, including their tacit knowledge, using robots. This initiative handles the foundational methods and technologies for the data that future robotics AI will learn, and we believe it will significantly contribute to robotics technology. ■ Career Path 【Specialist / Tech Lead in Recognition AI / Robotics AI】 You can build a career as a technical lead for projects, deepening your expertise in recognition AI, VLM, and VLA models. AI development knowledge based on a deep understanding of data will be an extremely valuable skill set globally in the future. 【Engineering Manager (EM)】 As the team expands in the future, you can aim for a leadership position responsible for managing, recruiting, and developing the engineering organization. Required Skills 【Development Experience】 ・Practical development experience using Python ・Development experience in a Linux environment, and team development experience using Git/GitHub ・Development and operation experience utilizing cloud platforms (AWS, GCP, Azure, etc.) 【Specialized Areas】 ・Experience analyzing image, video, point cloud, and other data using AI models, etc. ・Knowledge and experience in machine learning/deep learning 　・Experience building training, inference, and evaluation pipelines for models using frameworks such as PyTorch, TensorFlow, JAX, etc. 　・Theoretical knowledge of general statistics, machine learning, and computer vision 【Soft Skills and Stance】 ・An attitude of strong interest in the people and activities that generate data, not just the data and the technologies that handle it, and a desire to deeply understand the context surrounding the data. ・Communication skills to align technical requirements with internal and external stakeholders (PMs, clients, annotators, etc.) ・Problem-solving ability to independently derive solutions for uncharted technical challenges through literature research and hypothesis testing. Preferred Skills 【Tacit Knowledge / Embodied Knowledge】 ・Experience in research activities related to tacit knowledge and embodied knowledge. ・Experience in digitizing human movements and analyzing them while correlating them with linguistic information about the movements. 【Advanced AI / Model Development】 ・Experience implementing, training, and evaluating VLA (Vision-Language-Action) models, VLMs, and LLMs. ・Knowledge of efficient training and inference methods for large-scale models. ・Experience using libraries such as Transformers, PEFT (LoRA/QLoRA), Unsloth, vLLM. 【MLOps / Data】 ・Experience operating experiment management tools (MLflow, WandB, etc.) and workflow engines (Hydra, etc.). ・Experience handling robot datasets (Open X-Embodiment, etc.). Ideal Candidate Profile ・Individuals who deeply understand client challenges and can consider the social implementation of technology beyond just research and development. ・Individuals who can organize their own points of discussion and move forward by hands-on work, even on highly uncertain themes. ・Individuals who can proactively drive projects while involving others. ・Individuals with strong intellectual curiosity for new markets and technologies, and who can continue to learn. ・Individuals who resonate with FastLabel's purpose and can commit as stakeholders in industrial transformation. Application Overview Salary ■ Salary Annual salary: 6,000,000 - 15,000,000 JPY Monthly salary: 500,000 - 1,250,000 JPY ・Includes base salary (363,634 - 909,085 JPY) + fixed overtime allowance for 45 hours (125,234 - 313,085 JPY) + fixed late-night allowance for 20 hours (11,132 - 27,830 JPY). ・Compensation for hours exceeding the fixed amounts will be paid separately. *Consideration will be given to current annual income, and the final salary will be determined based on role, expected mission, and performance level. ■ Performance Review ・Twice a year (January/July) Work Location Tokyo Headquarters, Robot Training Center (Heiwajima), or client sites (Tokyo metropolitan area). Relocation may occur if new bases are established due to business expansion. ■ R&D Center Tokyo Ryutsu Center, Center Building, 6-1-1 Heiwajima, Ota-ku, Tokyo ■ Tokyo Headquarters Shinjuku Sumitomo Building 24F, 2-6-1 Nishi-Shinjuku, Shinjuku-ku, Tokyo Scope of Change: Generally not planned. Relocation may occur if new bases are established due to business expansion. Employment Type Full-time employee Work System ■ Work System Flextime system (5:00 - 22:00) *Core time is from 10:00 to 14:00. ■ Holidays and Leave Holidays: Complete two-day weekend system (Saturdays, Sundays, national holidays, New Year holidays, etc.) *Annual holidays of 124 days or more. Leave: Paid leave, congratulatory/condolence leave, menstrual leave, childcare/family care leave, etc. Probationary Period 3 months Benefits ・Various insurances complete (Health Insurance, Pension, Employment Insurance, Worker's Accident Insurance) ・Commuting allowance ・Qualification acquisition support system ・Referral recruitment support system (employee referral bonus, lunch expense subsidy)