Othman M. Benomar, PhD

Senior / Principal Data Scientist

Download PDF

Summary

Lead Data Scientist and ML engineer with 15+ years of experience building statistical models, anomaly detection systems, and production ML platforms across industry and astrophysics. His work covers Bayesian inference, deep learning, time-series modeling, and observability tooling, with a focus on clear evaluation in noisy, changing systems. He currently leads the ML stack behind PipelineML at Craftsman Software, turning first-principles research into production systems.

Experience

Lead Data Scientist

Craftsman Software, Tokyo, Japan

May 2024 to Present

  • Designed and deployed real-time ML systems for anomaly detection in distributed cloud environments.
  • Built CNN, LSTM, and attention-based models for sequential prediction and monitoring.
  • Built a Bayesian adaptive log anomaly detector around execution-path patterns extracted with Drain3.
  • Implemented RAG, MCP, and LLM pipelines for context-aware log and code analysis.
  • Developed MLOps workflows covering feature pipelines, model governance, and inference monitoring.

Project Associate Professor (Postdoc)

National Astronomical Observatory of Japan, Tokyo, Japan

July 2019 to June 2024

  • Directed statistical and ML research programs for multi-terabyte sequential datasets in HPC environments.
  • Developed forecasting, classification, anomaly detection, and inversion models.
  • Created data validation, interpretability, and uncertainty quantification pipelines.
  • Supervised PhD students and built internal ML training material.
  • Led collaborations with NASA, ESA, and partner research institutions.

Research Associate

New York University Abu Dhabi, Abu Dhabi, UAE

November 2015 to June 2019

  • Built large-scale statistical and ML pipelines for sequential and imaging satellite data.
  • Developed physical models with structures similar to ranking, retrieval, and personalization systems.
  • Worked with UAE stakeholders on data-driven research initiatives.
  • Supervised PhD students on time-series analysis projects.

Research Fellow

The University of Tokyo (JSPS), Tokyo, Japan

October 2013 to October 2015

  • Developed Bayesian inversion and latent-variable models for complex physical systems.
  • Secured a competitive ¥11M research grant.
  • Taught advanced statistical modeling for astrophysics graduate students.

Postdoctoral Fellow

The University of Sydney, Sydney, Australia

October 2010 to October 2013

  • Built high-dimensional Bayesian models for space-telescope telemetry.
  • Secured ARC funding and supervised research students.
  • Taught physics to undergraduate students.

Education

PhD, Applied Physics

Université Paris-Sud XI and École Polytechnique, France

Statistical methods applied to stellar plasma physics, with a focus on asteroseismology and Bayesian time-series analysis.

2010

MSc, Plasma and Optics Physics

Université Paris-Sud XI and École Polytechnique, France

Graduate training in plasma physics, optics, and diagnostic methods for industrial and astrophysical systems.

2007

Publications

Peer-reviewed astrophysics papers, conference talks, and seminar material are listed on the research page.

See full publication list →

Skills

Languages

  • Python
  • SQL
  • C++
  • Go
  • R

Frameworks

  • PyTorch
  • TensorFlow
  • Keras
  • Hugging Face
  • Dash

Tools

  • Kubernetes
  • Docker
  • MLflow
  • Prometheus
  • Grafana
  • Spark
  • AWS

Methods

  • Bayesian inference
  • Anomaly detection
  • Time-series modeling
  • Embeddings
  • Causal inference
  • Uncertainty quantification

Languages

  • French, native
  • English, fluent
  • Japanese, conversational
  • Moroccan Arabic, conversational