Selected work
Systems already in the field,
not slides about them.
A sample of what we ship: a simulation-grade suite for autonomous flight, a live operational platform for healthcare, and a computer vision pipeline running in production. Full case studies, screens, and stack on request.
Autonomous Vehicles · Simulation & Training
High-Fidelity 6DOF UAV Simulator
Devron built a high-fidelity 6-degree-of-freedom (6DOF) UAV simulator that models a real aircraft closely enough for the autopilot to never know the difference: a physics-based engine model, a full aerodynamic (aero) model, and an environment and sensor model covering wind, atmosphere, GPS, IMU, and camera inputs. It runs as two things at once. To an engineering team, it's a development rig for validating flight control algorithms against realistic dynamics long before hardware is involved. To a training organization, it's a full training simulator with dedicated instructor and student panels, letting an instructor inject faults, set scenarios, and monitor a trainee flying the same aircraft model in real time.
Simulation core6DOF engine, aero & environment/sensor models
Engineering useFlight control algorithm development & validation
Training useDual instructor & student panels, live scenario injection
6DOF UAV Simulator · dual flight stations
The simulator rig. Mission map and telemetry on the left displays, live cockpit view and engine/attitude gauges on the right, with dual flight control joysticks for instructor and student.
Custom Software Development · Healthcare
SmartQueue: Real-Time Hospital Queue Management
Hospitals lose time, and patient trust, to queues nobody can see. SmartQueue replaces the paper token board with a live, multi-role platform: patients track their exact position from their phone, receptionists issue and manage tokens in seconds, doctors run their consultation queue with one tap, and hospital administrators get a real-time, cross-department view of every queue on the platform. It's built as a true multi-tenant system, so one deployment serves every department, every hospital, every role.
Roles servedPatient · Receptionist · Doctor · Super Admin
Core mechanismLive WebSocket queue sync, sub-second updates
ScopeMulti-hospital, multi-department, role-based access
Receptionist · Register walk-in
Receptionist view. Token generation, walk-in registration, and full queue control from a single dashboard.
Doctor · Consultation session
Doctor view. Call, complete, pause, or transfer a consultation, with live on-time rate and session stats.
Super Admin · Platform overview
Admin view. Platform-wide visibility across every hospital, department, and queue in one console.
Patient · Queue tracker (mobile)
Patient view. Live token position, estimated wait, and doctor status, updating in real time on their own phone.
Artificial Intelligence & Machine Learning · Computer Vision
Vehicle Speed Detection: Real-Time Highway Monitoring
Devron built a vehicle speed detection system that turns an ordinary highway camera feed into calibrated, per-vehicle speed data. A custom-trained YOLOv26 model detects and classifies vehicles frame by frame, and a tracker follows each one across frames, measuring its pixel displacement between frames and converting that to real-world speed using a calibrated ground-plane mapping of the road. Vehicles are tagged with a persistent ID and a live speed reading, and anything over the configured limit is flagged for review automatically. The pipeline runs live on edge GPU hardware for highway monitoring, toll corridors, and traffic-enforcement deployments.
Detection modelCustom-trained YOLOv26 for vehicle detection & classification
Speed estimationFrame-to-frame tracking, calibrated to real-world distance
DeploymentReal-time inference on edge GPUs, automatic over-limit flagging
Vehicle speed detection · highway camera feed
Live detection feed. Each vehicle gets a persistent ID and a calibrated speed reading; anything over the configured limit is flagged automatically. Photo: real highway traffic, overlay is Devron's detection output.
