Weather Data Scientist

About the role

You will own the design and execution of data assimilation systems that power high-accuracy weather forecasts for the electric grid at Pravah. This role requires you to translate complex physical observations into reliable digital representations that inform operational decision-making. You will be responsible for ensuring observational integrity and geospatial consistency across massive, multi-source global datasets. You will directly contribute to the development of ML-ready pipelines that serve as the foundation for next-generation forecasting models. You will work at the cutting edge where physical modeling meets modern machine learning to solve real-world infrastructure challenges. Your work will support critical sectors such as renewable energy by providing precise solar and wind predictions. You will be a key contributor to systems that run at operational scale and must degrade gracefully under data stress.

Key facts

What you'll do

Build and operate a cycling data assimilation pipeline for our operational forecasting models, and produce the high-resolution gridded products it enables downstream.

Choose, deploy, and adapt a modern DA framework such as JEDI/UFO, GSI, DART, or PDAF to align with our regional and global forecasting requirements.

Develop and maintain observation quality control, bias correction VarBC, and thinning workflows that remain robust at high data volumes and degrade gracefully during feed outages.

Contribute to the development and integration of AI-based data assimilation pipelines that enhance traditional physics-based methods.

Tailor weather prediction models to renewable-sector needs, with a strong focus on solar (GHI) and wind generation (100m winds) parameters.

Assist in training and refining AI-based weather prediction models through careful curation of training data and validation against physical benchmarks.

Work at the intersection of physics-based modeling and machine learning, including hybrid physics-ML systems, learned parameterizations, and emulators.

Procure, process, and create ML-ready global and regional weather datasets at large scale, emphasizing coverage in data-sparse regions and long historical time horizons.

Ensure that all data processing steps are reproducible and production-grade, supporting reliable deployment in high-stakes operational environments.

Communicate technical methodologies clearly to both domain experts and cross-disciplinary teams to ensure alignment on objectives and outcomes.

Maintain and evolve assimilation workflows using a diverse suite of observational sources including satellite, radar, radiosonde, and station data.

Implement innovation statistics and diagnostic tools that provide insight into assimilation performance and observation impact.

Optimize data handling strategies for TB-scale, high-dimensional datasets including reanalysis, satellite, radar, and weather-station observations.

Manage geospatial transformations, grid definitions, and reprojection tasks to ensure compatibility across models and observation types.

Requirements

A master's or PhD in geophysical sciences, physics, applied mathematics, computer science, statistics, or a related field is required; a bachelor's degree with 3+ years of relevant research or operational experience is also acceptable.

You must demonstrate deep, hands-on expertise in data assimilation, proven through operational work, model contributions, research projects, publications, or technical reports.

You must have hands-on experience across the full data assimilation toolkit, including observation operators and error specification; variational (3D-/4D-Var) or ensemble (EnKF, LETKF, EDA) methods; cycling workflows and innovation statistics; and assimilation of satellite, radar, radiosonde, or station observations.

You must have hands-on experience with at least one operational data assimilation framework such as JEDI/UFO, GSI, DART, or PDAF, including building observation operators and forward models.

You must possess working knowledge of bias correction (VarBC), adaptive quality control, and gross-error rejection techniques as applied to large-scale weather data.

You must have a track record of contributing to or maintaining assimilation code, or holding responsibility in an operational or quasi-operational forecasting pipeline.

You must have experience working with TB-scale, high-dimensional observational and modeling datasets, including reanalysis, satellite, radar, weather-station, and sounding data, and the geospatial pipework around them such as grids, reprojection, and masks.

You must demonstrate fluency with widely used reference datasets such as ERA5, MERRA-2, IMDAA, IMERG/GPM, and GOES/INSAT/Himawari.

You must have practical experience on High Performance Computers (HPCs) and be comfortable managing large computational workflows.

You must show fluency in the modern geoscience Python stack, including xarray, dask, zarr, and netCDF, for data manipulation and analysis.

You must have experience building reproducible, production-grade pipelines that can withstand operational demands and strict validation requirements.

You must exhibit excellent written and verbal communication skills, with the ability to explain technical work to both domain experts and cross-disciplinary collaborators.

You must be comfortable working within a distributed team across India and the US, managing time zones and asynchronous collaboration effectively.

Nice to have

Prior work on projects specific to Indian geography is valued and may strengthen your application.

Familiarity with coupled earth-system models is considered a positive attribute for this role.

Experience with ensemble and probabilistic forecasting, regional downscaling, or subseasonal-to-seasonal (S2S) prediction is preferred.

Experience working with operational forecasting agencies such as IMD, NCMRWF, ECMWF, or NOAA is regarded as an advantage.

Familiarity with AI-based weather prediction models and data assimilation techniques is noted as beneficial.

Comfort using agentic AI tools to accelerate development is seen as a useful skill in this rapidly evolving environment.

Publications in relevant conferences or journals are encouraged but not required.

Practical notes

The team is distributed across India and the US, so expect a few hours of evening overlap with US Pacific Time on most workdays.

This is a full-time position with a standard work schedule aligned with team availability.

Candidates must be located in or willing to relocate to New Delhi for this role.

No external visa sponsorship is provided at this time.

Applications will be reviewed on a rolling basis until the role is filled, so early submission is encouraged.

Travel requirements are not applicable for this role.

The position demands a strong commitment to operational excellence and continuous learning in fast-moving technical domains.