ML Research Engineer - Weather Forecasting & Evaluation

WindBorne Systems; Redwood City (CA); Hybrid; USD 140,000 - 240,000; Full time

Benefits offered by this job

401(k)

Dental, health, and vision insurance

Unlimited PTO

Stock Option Plan

Office food and beverages

Job summary

WindBorne Systems is building a data‑driven weather intelligence platform by designing, manufacturing, and operating advanced weather balloons. We seek an evaluation scientist to lead rigorous comparisons between WeatherMesh and state‑of‑the‑art models, focusing on forecast quality and uncertainty.

You will build fast, reproducible evaluation pipelines, develop agentic tools, and communicate results through scorecards and visuals for researchers, leadership, customers, and partners.

Qualifications

• Strong scientific judgment and skepticism.
• Excellent experimental taste for evaluation questions.
• Systems thinking with focus on reusable infrastructure.
• Experience evaluating ML systems on geospatial or time-series data.
• Strong Python skills with PyTorch, NumPy, and pandas.
• Able to investigate ambiguous results and communicate conclusions clearly.

Responsibilities

• Evaluation strategy development with the Meteorology team to compare WeatherMesh against leading models.
• Build quick evaluations and durable, reproducible evaluation infrastructure.
• Improve evaluation tools, including agentic AI-based tooling for investigating forecasts.
• Communicate model performance with scorecards and visualizations for researchers, leadership, customers, and partners.

Skills

Scientific judgment

Experimental mindset

Systems thinking

ML evaluation

Python

PyTorch

NumPy

Pandas

xarray

Independent investigator

Tools

PyTorch

NumPy

Pandas

xarray

Job description

WindBorne Systems is building a data‑driven weather intelligence platform by designing, manufacturing, and operating advanced weather balloons. We seek an evaluation scientist to lead rigorous comparisons between WeatherMesh and state‑of‑the‑art models, focusing on forecast quality and uncertainty.

You will build fast, reproducible evaluation pipelines, develop agentic tools, and communicate results through scorecards and visuals for researchers, leadership, customers, and partners.