← Back to site Download as PDF (may be outdated) PROFESSIONAL EXPERIENCE
- Application engineer at Mollie, an Amsterdam-based payments fintech whose single-API platform processes online and in-person payments for businesses across 30+ EEA countries.
- As an early-stage data scientist and software engineer, played a key role in the 'zero-to-one' product development from its alpha stage and continued to drive the 'one-to-million' scaling phase for clients like Angelini Pharma, Ducati, Brembo, and Pfizer.
- Engineered and deployed the company's signature anomaly detection algorithm, enabling predictive maintenance that has saved clients millions by preventing machinery failures.
- Designed and built scalable MLOps pipelines for critical SaaS features, including ML-powered machinery diagnostics and an innovative 'Talk-with-Data' LLM-based analytics tool.
- Participated as an assistant for a joint research for Moreno & Crilly.
- Research covered topics on managerial and financial decision-making from temporal perspectives.
- Worked on data collection, EDA, and algorithmic analysis on text-based data related to various economic and financial news vs. financial activities such as short-selling, etc.
EDUCATION
- Completed remotely whilst working full-time, partially due to COVID-19, partially due to financial motives.
- Relevant coursework: Computational Data Analytics, Analytics Modeling, Deep Learning, Data & Visual Analytics, Simulation, etc.
- Thesis & Practicum: Statistical & empirical analysis to reverse-engineer micro-edge models running on a closed-source industrial-grade vibration sensor.
- Relevant coursework: Computer Science, Machine Learning, Statistics, Econometrics, Multivariate Calculus, Linear Algebra, etc.
- Thesis: A comprehensive analysis of missing imputation techniques and its performance on ML models. Created and delivered a benchmarking suite alongside diagnostic tools to judge the best-suited imputation method for a specific dataset & task.
SKILLS
- Languages: Japanese Native, English C2, Italian B2
- Programming Languages: Python, SQL, JavaScript/TypeScript, Rust, R
- Machine Learning Frameworks: Scikit-Learn, Tensorflow, PyTorch, PySpark
- Web Frameworks: FastAPI, Flask, Next.js
- BI Tools & Charting: Apache Superset, Streamlit, Apache Echarts, D3.js, Matplotlib, seaborn
- Tabular Data & ML Pipeline tools: Polars, Pandas, Numpy, Scipy, duckDB, Airflow, Celery, Hadoop, Kafka, OpenCV
- Databases: PostgreSQL, Google BigQuery, TimeScaleDB, SQLite3
- Miscellaneous: Git, Docker, Kubernetes, Linux, Terraform
PROJECTS & OTHERS
- Voluta: a VS Code extension for editing and resending cURL commands in place. Inspired by Firefox's Edit and Resend feature.
- cli_yutaro: a personal site that boots a full Alpine Linux instance in the browser via WebAssembly (v86). (code)
- cli-calendar: a Rust
cal clone with extra features like weekend and holiday highlighting. - AI Shopping List & Receipt Scanner: leverages a multimodal (visual) LLM tied with embedding vectorization to scan and match items on receipts to a shopping list. Personal project.
- Hotel Reservation Prediction Model: a novel, layered approach leveraging reservation time and check-in day to model the speed and acceleration of a hotel reservation, improving preceding models' performance by a non-trivial margin. Built freelance for a hotel consulting SaaS company.
- Other one-off projects:
- Excel Assistant: LLM-leveraged Excel editor for financial analysts, built with FastAPI (async SSE) and Next.js.
- Formula 1 Stint Optimizer: live-streaming stint optimizer using Kafka, Grafana, and FastAPI.
- Active member of Python Milano & PyData Milano since 2023, frequently participating in biweekly meetups & conferences.