Machine Learning Researcher

El Segundo, CA

About Rainmaker

Rainmaker is pioneering a modern cloud-seeding system to increase precipitation, improve water availability, and address severe-weather challenges. We combine atmospheric science, weather-resistant UAS, radar and satellite observations, numerical weather prediction, novel sensing systems, and sustainable seeding technologies to design, operate, and evaluate precipitation-enhancement programs.

Research at Rainmaker is attached directly to operations. Our scientists and engineers collect proprietary observations, deliberately intervene in atmospheric systems, evaluate the results, and use what they learn to improve the next operation.

About the Role

Rainmaker is hiring its first dedicated Machine Learning Researcher. You will establish how Rainmaker uses machine learning across the company: identifying the most valuable problems, determining which are ready for ML, building working models, and partnering with engineers and domain experts to turn successful research into operational systems.

You will not inherit a single predetermined model roadmap. The opportunity set includes forecasting supercooled liquid water and cloud-seeding opportunities, assimilating multimodal observations into estimates of atmospheric state, improving microwave-sounder retrievals, predicting hail, learning from intervention outcomes, and finding other high-leverage applications across research and operations.

Rainmaker's long-term advantage is not a generic weather model. It is the combination of proprietary in-cloud observations, radar and satellite data, UAS measurements, field campaigns, and repeated atmospheric interventions. You will build the learning systems that turn those data into better estimates, predictions, and decisions.

This is initially a hands-on individual-contributor role. You may eventually help recruit or technically lead an ML team if that fits your strengths and Rainmaker's needs, but management is not an initial expectation.

What You'll Do

Assess potential ML projects across Rainmaker and prioritize them by operational value, data readiness, technical tractability, and time to useful results.

Deliver an operationally useful model or prototype within your first three months rather than spending a quarter exclusively on infrastructure or roadmap development.

Build models for forecasting, nowcasting, retrievals, multimodal atmospheric-state estimation, simulation, intervention analysis, and other scientific or operational applications.

Develop methods for forecasting the occurrence, location, amount, and persistence of supercooled liquid water at scales relevant to cloud-seeding operations.

Combine public NWP, radar, satellite, microwave-sounder, aircraft, UAS, sounding, surface, and in-situ observations.

Build datasets, labels, baselines, evaluation metrics, and validation procedures for variables that public systems do not observe or optimize well.

Establish honest experimental comparisons and characterize calibration, uncertainty, generalization, and failure modes.

Work closely with meteorologists and atmospheric scientists to define targets, physical constraints, useful priors, and ground truth.

Write research-quality software and build prototypes that software engineers can help productionize when an approach proves valuable.

Use Rainmaker's compute budget deliberately, scaling experiments only when the problem, data, and baseline justify it.

Help Rainmaker learn from every operation, field campaign, new sensor, and intervention.

Communicate results and limitations clearly to scientists, engineers, operators, and company leadership.

What We're Looking For

Evidence of exceptional ability in machine learning research and engineering, regardless of whether it was developed in academia, industry, independent work, or another technical field.

Strong command of modern machine-learning methods and practical experience training, evaluating, and debugging models.

Strong Python skills and experience with a modern ML framework such as PyTorch, JAX, or an equivalent system.

Ability to turn ambiguous problems into measurable targets, tractable experiments, credible baselines, and working prototypes.

Sound statistical judgment, including careful treatment of leakage, distribution shift, calibration, uncertainty, and small or biased datasets.

Ability to work with noisy, sparse, multimodal, spatial, or temporal data.

Willingness to select simple methods when they are sufficient and reserve complex models for problems where they create measurable value.

Comfort working directly with scientists and engineers from domains you may not initially know.

High agency, rapid learning, and a strong bias toward useful results.

We care deeply about demonstrated technical ownership. If you have a project, system, experiment, paper, portfolio, or technical write-up that shows how you approach difficult problems, include it with your application and tell us what you personally contributed.

Preferred Qualifications

Experience with weather, climate, remote sensing, geospatial data, scientific ML, robotics, autonomy, aerospace, state estimation, computer vision, physical systems, or another data-constrained scientific domain.

Experience with forecasting, sequence models, probabilistic models, generative models, representation learning, sensor fusion, or data assimilation.

Experience working with radar, satellite, microwave-sounder, image, trajectory, gridded, or in-situ sensor data.

Experience taking a research model into real user workflows or production in partnership with software engineers.

Experience designing data-collection or labeling strategies when the existing dataset is insufficient.

Familiarity with atmospheric science is valuable but not required.

Starting Resources

Rainmaker will provide a dedicated compute budget, access to observations from its sensor fleet, growing proprietary datasets from operations and field campaigns, and close collaboration with atmospheric scientists and software engineers.

The data will not always arrive in a polished benchmark. Part of the role is determining what can be learned now, what ground truth must be improved, and which new observations would most increase future model performance.

What Success Looks Like

Within your first three months, you will have audited Rainmaker's most promising ML opportunities, selected a narrow and valuable initial problem, established a credible baseline, and delivered an operationally useful model or prototype with a concrete evaluation.

Within your first year, you will have established a prioritized ML roadmap grounded in actual data readiness and operational value; delivered one or more models that materially improve a scientific or operational workflow; and created reusable datasets, evaluations, or modeling foundations that accelerate subsequent work.