TRAVEE: VR neuro-rehabilitation platform
Software Developer with DevOps, ACS-UPB
Unity serious games fused with EEG, FES and haptic feedback, plus AI analytics, to accelerate post-stroke recovery in clinics and at home, with a full CI/CD and cloud backbone on AKS.
Problem
Stroke rehabilitation needs engaging, measurable, personalised exercise, and clinicians need remote visibility into session quality and adherence.
Solution
A VR platform of therapeutic mini-games driven by a state machine that adapts difficulty to live performance, synchronised with EEG (Emotiv/OpenBCI), FES stimulators and haptic gloves, backed by ASP.NET Core APIs and a containerised cloud backend.
My contribution
- Built Unity (C#) mini-games for reach, balance, gait and coordination with OpenXR, Quest and Windows MR support.
- Engineered data handlers synchronising EEG, FES and haptic streams into a state machine that auto-tunes difficulty from real-time performance metrics.
- Modularised the Unity codebase into packages/ScriptableObjects; exposed REST and WebSocket APIs (ASP.NET Core) for profiles, session logs and remote progress; OAuth2/SAML auth and AES-256 encryption.
- GitLab CI for PlayMode/EditMode tests, asset validation and multi-platform builds (Windows, Android, WebGL), scaled with Jenkins/TeamCity agents; Snyk + SonarQube gates.
- Dockerised micro-services deployed with Helm on AKS; Terraform (Azure & AWS), Vault/Ansible secrets; Prometheus/Grafana and ELK with Slack alerts.
Stack
Unity
C#
.NET
OpenXR
Python
GitLab CI/CD
Jenkins
TeamCity
NUnit
Snyk
SonarQube
Docker
Kubernetes
Helm
Terraform
Ansible
Microsoft Azure
AWS
HashiCorp Vault
Prometheus
Grafana
ELK Stack
Jira
ConfluenceLinks
VR and game development
Therapeutic mini-games in Unity (C#) (object reach, balance, gait training and coordination) with OpenXR, Oculus/Quest SDKs and Windows Mixed Reality support. Custom data-stream handlers for EEG sensors (Emotiv, OpenBCI), FES controllers and haptic gloves synchronise visual, audio and tactile feedback. State-machine-driven exercise workflows adjust difficulty dynamically from patient performance metrics.
Analytics
Movement-analysis algorithms (Kalman filtering, ML classifiers) for fatigue detection and motor-pattern recognition, developed with neuroscientists; telemetry dashboards (Unity Analytics, custom WebGL) so clinicians can monitor session quality, success rates and adherence.
Architecture and integrations
Modular Unity packages and ScriptableObjects for rapid prototyping of new therapy modules; REST and WebSocket APIs in ASP.NET Core for user profiles, session logging and remote progress tracking; OAuth2/SAML authentication and GDPR-compliant AES-256 encryption.
DevOps
- CI/CD: GitLab CI runs Unity PlayMode/EditMode tests, asset validation and platform builds (Windows, Android APK, WebGL); Jenkins and TeamCity agents for parallel builds and smoke tests.
- Containers: session manager, analytics API and notification service as Docker micro-services on AKS, released with Helm across dev/test/prod.
- IaC and secrets: Terraform modules for Azure and AWS (VPCs, Azure SQL / RDS, storage, load balancers); Ansible and HashiCorp Vault.
- Observability and security: Prometheus/Grafana health and performance alerting, ELK with Kibana dashboards; Snyk and SonarQube gates in CI; RBAC across Git repositories and Kubernetes namespaces.
