Rainmaker Fellow, Machine Learning

About the role

You will translate ambiguous atmospheric questions into concrete machine learning tasks that can be tested and measured. You will own a scoped workstream that directly supports Rainmaker's operational research priorities while collaborating closely with atmospheric scientists and engineers. Your daily work will involve collecting, curating, and diagnosing unusual atmospheric datasets that few organizations can access. You will design reproducible baselines, run controlled experiments, and document failure modes so that results remain credible across changing weather regimes. You will work with real sensor and operational data, ensuring that models respect physical constraints and practical decision-making needs. You will communicate your findings clearly to both technical and non-technical audiences, including scientists, operators, and leadership. Ultimately, you will leave behind a durable dataset, model, or prototype that Rainmaker can continue using and building upon long after your fellowship ends.

Key facts

What you'll do

Establish machine learning pipelines around Rainmaker's unusual atmospheric datasets and operational requirements.

Build and improve training and validation datasets from noisy, multimodal, and spatially distributed weather observations.

Create reproducible baselines that isolate signal from noise before introducing complex modeling approaches.

Train, evaluate, and debug models across diverse weather events, regions, and operating conditions.

Quantify calibration, uncertainty, generalization, and failure modes while accounting for missing or biased data.

Collaborate with atmospheric scientists to define target variables, ground truth, physical constraints, and operational success metrics.

Write clear, reusable code and documentation that enable other team members to build on your work quickly.

Present results to Rainmaker's scientists, engineers, operators, and technical leadership in accessible formats.

Deliver a final artifact such as a benchmark dataset, evaluated model, prototype product, evaluation report, or research-ready analysis.

Support the selection and prioritization of future data collection and sensing strategies based on model performance needs.

Contribute to internal reviews, experiment tracking, and knowledge sharing across the ML and atmospheric science teams.

Explore opportunities to integrate remote sensing, radar, satellite, and aircraft-based observations into scalable ML workflows.

Assist in framing operational problems as machine learning problems without losing fidelity to real-world constraints.

Iterate quickly on model behavior while maintaining scientific rigor and traceability of assumptions.

Help maintain and version datasets and modeling artifacts so that long-term comparisons remain possible.

Requirements

Current undergraduate, master's, or PhD students; postdoctoral researchers; recent graduates; and other early-career researchers are all eligible.

Strong Python programming ability and experience with a modern machine learning framework.

Demonstrated capacity to independently build, test, and debug technical work from data to model.

Strong quantitative reasoning and an ability to design credible experiments.

Comfort working with noisy, sparse, multimodal, spatial, temporal, or physical data.

Proven ability to make progress on ambiguous research problems while incorporating mentor feedback.

Clear written and verbal communication skills suitable for technical and non-technical audiences.

Availability for full-time, on-site work in El Segundo for the agreed appointment.

U.S. work authorization is required for this role.

Eligibility for benefits may be subject to specific employment classification and local regulations.

Nice to have

Prior experience with forecasting, sequence modeling, computer vision, state estimation, sensor fusion, probabilistic modeling, data assimilation, or uncertainty quantification.

Background in weather knowledge, remote sensing, robotics, autonomy, geospatial analysis, or scientific computing.

Experience contributing to open source or maintaining reproducible data science workflows.

Familiarity with atmospheric observations, radar, satellite data, or numerical weather prediction.

Practical notes

This is a paid, full-time internship based in El Segundo, California.

Appointments are three to six months in duration, with four months as the standard length.

Applications are reviewed on a rolling basis and start dates are flexible based on project and mentor readiness.

Compensation is $8,000 per month, with full health coverage (medical, dental, and vision insurance), lunch provided when working in-office, and a fully stocked kitchenette.

Free electric vehicle charging is available at the HQ.

Work authorization is required, and eligibility for benefits may be subject to employment classification and local rules.