Looking for a job

Taymaz Ghaneh

@taymazghaneh

Message

Data Scientist at Boskalis who developed an LSTM/CRF pipeline to classify vessel activity from irregular AIS data.

Netherlands

Message

At Boskalis, I developed a reproducible machine-learning pipeline using LSTM/CRF models to classify vessel activity from irregular AIS data.

I evaluated model performance and reliability, quantified data quality, and analyzed errors. The models improved on baselines across two operationally relevant metrics, and I communicated findings and limitations to stakeholders.

For an independent financial time-series project, I built an end-to-end ML platform on Databricks, with training and inference orchestration, model serving, and a dashboard.

At Leiden University, I studied Statistics & Data Science. My projects included causal inference on smoking and hypertension and a memory-efficient MinHash/LSH pipeline for finding similar users in Netflix rating data.

Experience

Work history, roles, and key accomplishments

Graduation Internship in Data Science

Boskalis

Feb 2026 - Aug 2026 (6 months)

Developed a reproducible ML pipeline (LSTM/CRF) for vessel-activity classification using irregular AIS data, including data-quality quantification, model evaluation, reliability assessment, and error analysis. Achieved statistically significant improvements across two operational metrics and improved generalizability, communicating findings to stakeholders.

Python Databricks Spark SQL TensorFlow MLFlow LSTM CRF Machine Learning Data Quality Model Evaluation

Education

Degrees, certifications, and relevant coursework

University of Tabriz

Doctor of Philosophy, Gravitational Physics

Grade: 18.60 / 20

Ph.D. in Gravitational Physics from the University of Tabriz.

Leiden University

Master of Science, Statistics & Data Science

2024 - 2026

Grade: 7.81 / 10

MSc in Statistics & Data Science from Leiden University, with experience in dense time-series modeling and additional training in longitudinal modeling, causal inference, and Bayesian modeling.