July 2025 — December 2025
Laravel backend internship
Backend internship focused on Laravel development over a 6-month period, contributing to application features and production-oriented engineering work.
About
Samuel Jaglale — Software Engineer in La Réunion, France.
I work as a generalist. In practice that means the same week can include a training loop in PyTorch, a Laravel controller, and a Caddy config on a small VPS — and that the interesting problems usually sit where those things meet.
The through-line is preferring things that actually run. A reinforcement learning agent is only interesting once it clears a threshold across several seeds; an automation workflow is only interesting once it delivers email on a schedule without supervision. I try to build to that bar and to be precise about where a project currently stops.
Computer vision runs alongside this: OpenCV-based experimentation, camera workflows with an OAK-D Lite, and applied object recognition using YOLO in hackathon contexts.
01
Applied AI
02
Computer Vision
03
Backend Development
04
Analytical Systems
Trajectory
How the range was built.
Not a CV — a sequence of capabilities, each one attached to the work that proved it. Hover an entry to move along the spine.
Selected experiments
Work that did not become a case study.
Shorter pieces — hackathons and ongoing experimentation. Listed because they are real, not padded into something they were not.
Network eligibility and antenna scoring
HackathonGeospatial work on 4G and 5G coverage: GeoJSON handling, a map view of coverage, an eligibility API answering whether a given location is covered, and a scoring system ranking candidate antenna placements against terrain, remoteness, road access and population density.
GeoJSONMappingREST APIScoringFood recognition with YOLO
HackathonApplied object recognition on food images, covering dataset handling, inference and the gap between benchmark accuracy and what the model does on photographs taken in a real room.
YOLOOpenCVPythonCamera workflows with OAK-D Lite
Ongoing experimentationDepth-camera capture and on-device inference experiments, used mainly to understand where a vision pipeline actually spends its time.
OAK-D LiteDepthOpenCV
Stack
Tools I can actually defend.
Grouped by what the work needs rather than by logo count — each item here has been used on something that shipped or produced a result.
Applied AI & vision
Tools and workflows used for experimentation, inference and image-based recognition.
PythonPyTorchGymnasiumOpenCVYOLOOAK-D LiteNumPy
Backend & product
Web and application work backed by practical implementation choices.
Laravel / PHPTypeScriptNext.jsFlutter / Dartn8nPostgreSQLMySQL
Systems & analysis
Lower-level work where numerical behaviour and memory discipline actually matter.
CNumerical simulationGeometryCLI tooling
Deployment & ops
Operational habits that keep small systems usable and secure.
LinuxDockerDigitalOceanCaddySSH hardeningUFWGit
Contact
Have something that needs building?
Open to internships, freelance work and collaborations across applied AI, backend development and deployment.