Expertise
Operational air-quality systems that bring ground monitors, satellite products and chemical transport model output into one place decision-makers can use.
From December 2022 to September 2024 I was Air Quality Application Development Associate at ICIMOD. I supported the operation and maintenance of the air quality monitoring systems for the Hindu Kush Himalaya (HKH) region, the Government of Nepal and Bangladesh. In that role I designed and deployed the air quality database and the decision-support dashboard. I also wrote the automation that ingests ground observations, satellite products and model outputs and runs QA/QC on them. Alongside that, I worked on the HPC cluster that runs the WRF-Chem atmospheric model and the HYSPLIT dust forecast, supporting model execution, monitoring and the resolution of runtime errors.
Most air-quality programmes in the region don't fail on the science. They fail on the plumbing: forecasts that stop running quietly, station feeds full of gaps and spikes, satellite and model grids that never get put next to the observations. My work covers that layer. I build reproducible Python workflows for NetCDF, HDF and GRIB data, write database and service designs that operations teams can maintain, and build dashboards that a ministry officer can read. I work with the scientists who own the models and the methods, not in place of them.
Design, development and operational maintenance of AQ dashboards that integrate ground monitoring stations, satellite products and model output, as delivered for the HKH region, Nepal and Bangladesh at ICIMOD.
Running, monitoring and troubleshooting WRF-Chem simulations and HYSPLIT dust forecasts on Linux HPC clusters, including batch job submission, runtime-error resolution and post-processing of output for evaluation and display.
Automated ingestion, processing and quality control of station time series and gridded atmospheric data (NetCDF, HDF, GRIB) with Python, NumPy, Pandas and xarray, plus CDO and NCO where they fit better.
PostgreSQL/PostGIS schemas for monitoring networks, and standards-based delivery of gridded fields through THREDDS Data Server, ncWMS and GeoServer, so that dashboards and analysts read from a single trusted source.
Bringing emission inventory datasets into modelling and visualisation workflows: format conversion, regridding and display alongside concentrations. This builds on formal training in compiling air pollution emission inventories.
Working between model developers, international technical partners (SERVIR Science Coordination Office, NASA Applied Sciences teams) and government users in Regional Member Countries, including workshops, training and stakeholder demonstrations.
A maintained AQ platform with one database and one dashboard for station observations, satellite products and model forecasts, built for an agency to run itself.
8-16 weeks · fixed-scope project
Keep air quality and dust forecasts running reliably on HPC: automation, monitoring, troubleshooting and post-processing for display and evaluation.
3-8 weeks, or monthly retainer · fixed-scope project | retainer
Clean, documented, reproducible air quality datasets from station networks, satellites and emission inventories, plus training so your team can keep them that way.
2-6 weeks · fixed-scope project | advisory
ICIMOD · 2022–2024
Operational AQ monitoring systems for the HKH region, the Government of Nepal and Bangladesh, combining ground stations, satellite products and model output in one database and dashboard.
Countries: Nepal, Bangladesh, Bhutan
ICIMOD · 2022–2024
Supporting HPC runs of the WRF-Chem atmospheric chemistry model and the HYSPLIT dust forecast whose outputs fed ICIMOD's operational air quality dashboards.
Countries: Nepal, Bangladesh
Where
Loading map…
Tell me what you're working on. A 20-minute call is free and there's no obligation.