Bilingual Senior Data Scientist / ML Engineer – AI Marketing Tech, Hybrid (¥8M - ¥14M)

United World is a recruitment agency supporting global talent in Japan. We are hiring on behalf of our client, a fast-growing AI and marketing-technology company.

About the Company

A Series B start-up (approx. 51–100 employees) building AI-powered customer communication products used by brands across Japan, with offices in Tokyo and Osaka plus 3 overseas bases. The team includes members from 20 countries, and the Data & AI team works closely with AI engineers in Japan, India, Taiwan and the Philippines.

The Role

Senior Data Scientist / Machine Learning Engineer in the Data & AI team. You will lead the full ML lifecycle – from data exploration to production – on the company's own large-scale first-party data.

- Build and run ML models for personalised message delivery: timing and content optimisation, CLV prediction, customer segmentation

- Design and implement recommendation engines (rule-based, ML and deep learning)

- Lead clustering-based customer behaviour analysis

- Clean, validate and explore large structured and unstructured datasets

- Define business KPIs with sales, customer success and client marketing teams, and present data-driven proposals (e.g. expected ROI of a recommendation feature)

- Improve model accuracy through error analysis and iteration; manage production-grade code on GitHub

- Research and apply deep learning, NLP and LLM / generative AI

- Mentor junior members and set development standards

Requirements

- 4+ years of hands-on data science experience

- Strong ML and statistics: prediction, classification, regression, clustering on tabular and time-series data (NumPy, scikit-learn, PyTorch etc.)

- Python and SQL with attention to code quality

- Offline evaluation (cross-validation), A/B testing, basic statistical analysis

- End-to-end ML experience: feature engineering, KPI definition, training, deployment and operation in production

- Linux, Git/GitHub, Docker, CI/CD (e.g. GitHub Actions); experience shipping models with engineering teams

- Japanese: business level or above (meetings, writing and reading)

- English: comfortable discussing technical topics

Nice to Have

- Data warehouses / OLAP (Snowflake, Redshift), AWS

- Recommender systems (collaborative filtering, matrix factorisation, NCF, Two-Tower)

- MLOps, online serving, FastAPI/Flask microservices, fine-tuning open-source models, low-latency ML APIs

- Airflow / Step Functions, Kafka / Kinesis, Spark

- Ranking, search, ad delivery or personalisation projects

- OSS contributions, papers or Kaggle results

Conditions

- Location: Tokyo (Aoyama) or Osaka (Umeda) – your choice

- Hybrid work (about 1 office day per week recommended), flextime (core 10:00–16:00)

- Salary: JPY 8.04M – 14M per year (includes fixed overtime pay for 45 hours/month; extra overtime paid separately). Higher may be considered for strong candidates

- Full-time, permanent; 3-month probation with the same conditions

- Weekends and public holidays off, summer and year-end holidays, paid leave

- Book and seminar support (JPY 55,000/year), health check, commuting allowance

Selection Process

Document screening → Online coding test → 1st interview (English, technical) → Coding challenge (take-home) → 2nd interview (Japanese) → Final interview (English)