Shared technical foundation
All 8 team members – across software engineering, data management, and geospatial roles – now operate from a common cloud-native baseline, enabling faster, more precise technical decision-making.
A tailored hybrid training programme helped AODN unify technical understanding across its data team, and turn that alignment into real improvements in cloud-native data pipelines.
Across engineering, data management, and geospatial roles within AODN.
From STAC and data discovery to formats, processing, and applied workflows.
Described as one of the best technical courses attended by participants.
The Challenge
AODN needed to move beyond fragmented knowledge and build a shared foundation for cloud-native geospatial across its data team.
The Australian Ocean Data Network (AODN) of Australia’s Integrated Marine Observing System (IMOS) sits at the heart of Australia's marine data ecosystem, serving researchers, scientists, and government agencies with terabytes of ocean and climate data collected from a range of platforms, including satellites, ships, moorings, and autonomous vehicles.
As data volumes grew and user demands evolved, the AODN team has a growing need to transition data production pipelines to cloud-optimised formats like Zarr and Parquet and adopting cloud-native workflows at scale.
But modernisation requires more than tooling, it requires a shared language. The AODN team spanned a range of roles: software engineers, data managers, and geospatial data engineers. Each brought deep domain expertise, yet the team lacked a unified baseline understanding of cloud-native geospatial concepts. Without it, conversations around workflow design, format selection, and optimisation strategies were difficult to anchor.
A key specific gap was around chunking and partitioning; the architectural decisions that determine how data is stored, queried, and served. The team needed clarity on best practices and their downstream impact on performance, but there was no structured resource to bring everyone to the same level quickly and practically.
thriveGEO's Cloud-Native Geospatial 101 is a unique offering - exactly what we needed to bring everyone to the same level and leave them ready to act.
Why this mattered
The AODN team needed a shared understanding of cloud-native geospatial concepts to support the transition of their data pipelines to formats like Zarr and Parquet – enabling more consistent decisions around data structure, performance, and scalability.
The Solution
thriveGEO designed and delivered a bespoke two-week hybrid training programme – a carefully paced blend of self-directed online learning modules and live instructor-led practical sessions. Rather than using generic datasets, thriveGEO adapted the practical exercises to AODN's own data, ensuring that participants were learning new concepts through a lens they already understood.
A progressive curriculum guided participants from core concepts to applied cloud-native workflows.
Hands-on modules covered STAC, cloud-optimised formats, and key concepts like chunking and partitioning.
Participants worked with tools like xarray, Dask, and DuckDB to explore real-world data processing.
Each participant applied their learning directly to AODN datasets – bridging theory and operational impact.
The Results
The training achieved its primary goal: the entire AODN team now shares a common, robust understanding of cloud-native geospatial concepts. That alignment has already begun to pay dividends. With a shared vocabulary and a common conceptual framework, the team can now have more productive, more precise internal conversations about how to evolve AODN's data products.
Participants built practical experience with STAC, Zarr, GeoParquet, xarray + Dask, and DuckDB – applying these technologies directly to AODN’s own datasets.
The training surfaced concrete opportunities to improve AODN’s data pipelines, including a redesign of Parquet partitioning to better reflect how data is accessed and queried.
Participant Voices
Feedback highlighted both the accessibility of the teaching and the value of applying concepts through hands-on work.
I think it was one of the best, if not the best, technical course I've ever attended.
The team is already having richer, more informed conversations about how to make our products better and our processing pipelines more performant.
Business Impact
By aligning a cross-functional data team around a shared cloud-native geospatial foundation, AODN created the conditions to modernise its data pipelines with greater speed, clarity, and confidence.
All 8 team members – across software engineering, data management, and geospatial roles – now operate from a common cloud-native baseline, enabling faster, more precise technical decision-making.
The training directly triggered a review of AODN’s Parquet partitioning strategy – now an active workstream to improve query performance and downstream data delivery.
With hands-on experience across Zarr, Parquet, STAC, xarray + Dask, and DuckDB, the team is already identifying and implementing improvements to existing data pipelines.
Aligning the full team eliminated siloed understanding, reduced technical bottlenecks, and strengthened organisational resilience as cloud-native approaches become central to AODN’s roadmap.
Build geospatial capability
thriveGEO designs tailored learning programmes that build confidence, shared language, and practical capability across technical and cross-functional teams.
Subscribe to our monthly geospatial skill signal – with practical tools, workflows, and insights.
© 2026 thriveGEO GmbH | Imprint | Data Privacy