Surveillance Platform
Architected and deployed a production-ready surveillance platform supporting 10,000 CCTV cameras, 24/7 monitoring, and real-time incident management, independently deployed and configured for 3 government agencies.
Each government agency was managing surveillance operations through fragmented, outdated systems with no centralized way to monitor camera feeds, track incidents, or manage access within their own organization. The engineering challenge was building a single production-ready platform that could be deployed and configured independently for each agency, supporting nearly 10,000 camera feeds, strict access control, real-time incident monitoring, and 24/7 high-availability operations.
“Each agency needed a production-grade surveillance system built to their own security and operational requirements.”
Led full-stack engineering across frontend, backend, middleware, and infrastructure for a large-scale government surveillance platform. Owned architecture decisions, VMS integration, real-time monitoring workflows, and engineering standards across the platform.
Frontend architecture for real-time monitoring
Architected the frontend using Next.js, Redux Toolkit, and RTK Query to support live camera monitoring, incident management, and real-time dashboard updates across large-scale surveillance operations.
Backend and middleware services for VMS integration
Built NestJS APIs and secured middleware services that connected directly with third-party VMS systems inside protected intranet environments, handling camera sync, live streams, snapshots, and health status.
Geospatial map monitoring with Redis indexing
Built Mapbox monitoring features integrated with Redis geospatial indexing and real-time camera updates, improving operator visibility and incident response across distributed camera sites.
Multi-tier role-based access control and SSO
Designed a multi-tier permission system with OAuth 2.0 SSO, ensuring operators only accessed camera feeds and incident data within their assigned boundary.
Background processing for evidence and video clips
Built BullMQ worker pipelines for video clip generation, evidence processing, and MinIO S3 uploads, isolating CPU-intensive workloads from real-time monitoring services.
Monorepo setup and engineering standards
Introduced Lerna monorepo, automated testing with Jest and Cypress, and CI/CD pipelines via GitHub Actions across frontend, backend, and middleware services to improve code quality, reduce deployment risk, and keep engineering practices consistent across the team.
Infrastructure scaling and production delivery
Collaborated on deployment architecture using Docker and Nginx, supporting horizontal scaling, security hardening, and high-availability deployments across distributed government environments.
10,000+
CCTV Cameras Monitored
Supported across a single platform built for large-scale surveillance operations
3
Government Agencies
Platform independently deployed and configured for each agency
50%
Faster Map Rendering
Improved through optimized state management and geospatial indexing
~60%
Fewer Production Defects
Reduced through automated testing and TDD practices
20%
Faster Deployments
Achieved through CI/CD automation and monorepo pipeline improvements
24/7
Continuous Monitoring
High-availability surveillance operations across all agency sites
- 1
Clear service boundaries make systems easier to reason about. Separating frontend, backend, and middleware into distinct layers kept failures contained and made the system significantly easier to debug and scale as complexity grew.
- 2
State consistency is a design requirement, not an afterthought. Working on real-time monitoring taught me to treat synchronization between camera status, alerts, and frontend state as a core architectural concern from the start.
- 3
Offloading heavy processing early keeps critical workflows stable. Moving video clip generation into background workers before it became a bottleneck meant live monitoring stayed unaffected even during the heaviest alert spikes in production.
- 4
Government delivery sharpens your instinct for reliability. Operating in a mission-critical environment shifted my priorities toward deployment stability, fast production support, and recovery planning as core engineering responsibilities.
- 5
Security and compliance are engineering constraints, not checklists. Deploying on government infrastructure taught me to treat security hardening and GDPR compliance as first-class requirements built into the architecture, not steps added at the end.