PaulB_
Senior Member
What started as a lifelong personal interest in media technology and problem-solving in my bedroom naturally evolved into a massive, production-grade undertaking. One thing led to another as I pushed past limits to see what could actually be built.
Running my independent IT practice—handling remote and on-site system administration, enterprise hardware repairs, complex data migrations, and advanced network troubleshooting—has sharpened my operational discipline. Leveraging scalable, entry-level Netcup cloud instances hosted from my setup, I engineered, deployed, and currently operate a fully accessible, multi-region streaming ecosystem. Whether I am based in Cape Town, on-site in Johannesburg, or visiting family in New Zealand, this infrastructure remains globally accessible and rock-solid.
I built this environment entirely from the ground up out of a passion for tackling complex challenges and engineering custom solutions where standard software (the closest commercial alternative being something like Plex) fell short.
I'm testing the waters here to see what opportunities exist for integration or architecture work, and to connect with technical leads who understand complex distributed media systems. Here is the actual technical scope of what I built and run:
Based in Cape Town, available for remote gigs or on-site in JHB if needed. Source code stays private to protect the setup, but happy to jump on a demo or show screenshots to anyone who actually gets how deep this stuff goes. Hit me up via PM if you want to chat.
Some Screenshots
Running my independent IT practice—handling remote and on-site system administration, enterprise hardware repairs, complex data migrations, and advanced network troubleshooting—has sharpened my operational discipline. Leveraging scalable, entry-level Netcup cloud instances hosted from my setup, I engineered, deployed, and currently operate a fully accessible, multi-region streaming ecosystem. Whether I am based in Cape Town, on-site in Johannesburg, or visiting family in New Zealand, this infrastructure remains globally accessible and rock-solid.
I built this environment entirely from the ground up out of a passion for tackling complex challenges and engineering custom solutions where standard software (the closest commercial alternative being something like Plex) fell short.
I'm testing the waters here to see what opportunities exist for integration or architecture work, and to connect with technical leads who understand complex distributed media systems. Here is the actual technical scope of what I built and run:
1. Cost-Engineered Storage & High-Availability Disaster Recovery
- Zero-Cost 165TB Backend Architecture: Architected, deployed, and stabilized a massive 165TB media storage backend by integrating and configuring alternative high-capacity storage pipelines, proving that petabyte-scale storage can be engineered with extreme cost-efficiency.
- Redundant Business Continuity & Mirroring: Implemented a robust, automated failover and mirroring strategy utilizing OpenDrive to protect mission-critical assets against sudden vendor outages or service deprecations.
- High-Stakes Crisis Management & Recovery: Following a catastrophic vendor outage where account access was abruptly revoked—leaving me staring down the total loss of my library and initiating a complete rebuild—I successfully navigated the crisis until access was restored, executing a high-stakes emergency 5TB data recovery upon service restoration with zero data loss.
- Bypassing FUSE Bottlenecks for Direct HTTP Streaming: Standard FUSE mounts introduced severe I/O latency and streaming stutter across the massive storage array; I engineered a direct HTTP streaming architecture linking the core application straight to rclone, entirely bypassing FUSE overhead and completely eliminating buffering issues.
- Mount Resilience & Filesystem Protection: Engineered rigorous defensive scripts and system safeguards—including strict os.path.ismount validation checks and isolated staging directories (/var/lib/pvr/staging), preventing active media writes from dumping onto and corrupting the root filesystem if a remote mount drops mid-recording.
2. Massive Data Curation & Ingestion Pipeline
- Large-Scale Data Ingestion & Normalization: Personally ingested, manually curated, and standardized over 20,000+ music videos through iterative trial and error, establishing a pristine, uniform media library free of stream errors or aspect ratio mismatches.
- Rigorous Metadata Architecture: Built a meticulous metadata tagging and structuring pipeline, hand-cataloging artist names, track titles, release years, and genre classifications to guarantee flawless automated scheduling and consistent playback across custom channels without relying on brittle, inaccurate automated scrapers.
3. Proprietary On-Demand Playout Engine
- Extreme Resource Optimization: Custom-coded a lightweight, high-performance streaming engine optimized specifically for constrained hardware resources to handle continuous multi-channel distribution.
- Strict On-Demand Lifecycle Management: Designed the core streaming architecture to be strictly on-demand, ensuring linear broadcast streams and channels only spin up and execute compute cycles when a client actively connects, instantly freeing up CPU, RAM, and network bandwidth when inactive.
- Proxy Fan-Up & Upstream Constraints Handling: Engineered custom proxy middleware and fan-out logic to navigate single-tuner provider constraints, ensuring smooth stream injection, connection multiplexing, and stable multi-client handling without dropping upstream connections.
4. Concurrent Dual-Broadcasting Network (Video & Audio)
- ChannelWorks (Video Headend): Architected and deployed a multi-channel linear video broadcasting network equipped with automated round-the-clock scheduling, time-zone synchronization, and broadcast logic.
- Breathe Radio & AudioWorks Master Console: Built an entirely parallel, concurrent audio broadcasting network featuring advanced dual-deck management, live running order synchronization, and real-time audio stream orchestration running simultaneously alongside the video headend.
5. Broadcast-Quality Graphics & Real-Time OSD (On-Screen Display)
- Native OSD Layering Engine: Designed clean, broadcast-grade lower-thirds and dynamic UI overlays—rendering network bugs, artist tags, track titles, and release years in real time over active 4K/HEVC video streams.
- Live News, Weather & Audio Ducking Integration: Embedded real-time regional news and weather data directly into the custom OSD layout pipeline while simultaneously streaming and ducking Breathe Radio audio tracks in the background.
- Zero-Lag UI Performance: Optimized rendering pipelines to ensure high-density graphical elements overlay cleanly onto live video feeds with sub-millisecond precision.
6. Custom Android TV Client & Rich Media Hub (Kotlin / Jetpack Compose)
- Bespoke Set-Top-Box Experience: Developed a customized, professional Android TV application from scratch using Jetpack Compose, inspired by high-end broadcast boxes like Enigma2 and Slyk Q.
- Advanced Player Lifecycle Integration: Engineered robust surface holder lifecycles and seamless dynamic player switching between libVLC and ExoPlayer, resolving complex hardware acceleration, buffer allocation, and texture rendering hurdles.
- Immersive Music & On-Demand Hub: Built deep media navigation supporting hierarchical album organization, artist showcases, synced lyrics displays, interactive audio visualizers, and detailed metadata tracking.
- Secure API & Headless Pairing Topologies: Designed secure backend integrations featuring a short-code device pairing workflow (/api/device-pairings) for headless TV authentication, paired with Netmaker overlay mesh networking to secure all server-to-server communications.
Based in Cape Town, available for remote gigs or on-site in JHB if needed. Source code stays private to protect the setup, but happy to jump on a demo or show screenshots to anyone who actually gets how deep this stuff goes. Hit me up via PM if you want to chat.
Some Screenshots