Jaguar Land Rover · Platform & partnerships
Build for the platform. Not the partner.
Protecting a reusable vehicle-data platform while agreeing architecture and delivery commitments with NVIDIA.
Four vehicle-data use cases. One reusable platform capability.
01 / problem
The problem to solve
Large-file transfer needed to support multiple vehicle-data products, not a single integration. A partner-specific approach risked narrowing the platform, duplicating future capability and shifting additional AWS operating cost onto JLR.
02 / decision
The decision
I chose to protect Large File Transfer as reusable platform infrastructure rather than optimise it around one partner. I challenged the proposed cost model and made retrieval ownership explicit, keeping immediate delivery commitments within a sustainable longer-term product direction.
03 / approach
How I approached it
I worked directly with NVIDIA and with Engineering, Cloud and Architecture colleagues to turn external commitments into an achievable internal roadmap. We evaluated retrieval and cost models against partner-side storage-access charges, while sequencing delivery across competing use cases and platform dependencies.
04 / outcome
What changed
The resulting capability supported four distinct connected-vehicle data use cases on one platform: uploading diagnostic logs for analysis, centrally managing vehicle logging configuration and collecting the results, deploying data-recording campaigns to targeted vehicles and retrieving what they capture, and collecting large-scale mapping and ADAS datasets. Vehicle-data offboarding was delivered while preserving a reusable foundation for subsequent products, rather than a one-off, partner-specific build.
05 / What I take forward
What I take forward
Vehicle-data offboarding at scale is not free, so cost was part of the product decision from the start. I built retrieval and storage-cost modelling into the design, not just delivery scope, and used it to sequence which use cases went first.
Explore another story