Senior Fullstack Cloud Engineer, Java / Angular
May 2026 to PresentMapersive · Immersive spatial-data platform · Casablanca
- Sped up loading on a data-heavy global map by slimming the payload the backend returns to only the fields the map renders, and fetching just the features inside a buffer zone around the user's current view rather than the whole global dataset.
- Integrated Potree point clouds into the platform end to end: dense photogrammetry clouds are converted into level-of-detail octrees by a Dockerized PotreeConverter job on AWS Batch, tracked through SQS status events, and streamed to the viewer through presigned range requests, so very large clouds render over the panoramas fast and reliably instead of being downloaded whole.
- Resolved a production outage caused by heap exhaustion: nested JPA projections were loading whole entity graphs, including geometry and eagerly loaded relations, for every row. I replaced them with flat projections, so the listing now runs as a single query of 13 columns.
- Stopped a map listing endpoint from loading full entities: 6 entities per row down to 0, with related geometry fetched in exactly one query and protected by regression tests.
- Built point cloud measurement tools in Angular (point, distance, height, vertical and horizontal surface) that convert local coordinates to WGS84, so results are stored as real geographic positions.
- Added background prefetching of the point cloud octree so it streams in before the user opens it, plus a point-budget control from 500k to 8M points.
- Run sprint planning and retrospectives, breaking client requirements into scoped, prioritized work and keeping delivery on track.
- Work directly with Moroccan and overseas clients, turning requests into concrete engineering tasks and committed timelines; delivered their full feature requests on short, committed timelines.
- Improved the visual quality of the KRPano viewer and built a report on top of the platform's photogrammetry output.
- Java
- Spring Boot
- JPA
- Angular
- AWS Batch
- SQS
- Potree
- KRPano