01AI & Local Model Infrastructure
Private inference infrastructure, owned outright rather than rented.
- Local models across multi-GPU servers
- Quantization, VRAM and context planning
- Task-based routing between models
- Vision, OCR, speech and multimodal pipelines
02Real-Time Media & Streaming
Live video and audio treated as an engineering problem with a latency budget.
- One live input, many destinations
- OBS, RTMP, media servers and VOD
- Transcription, OCR and event detection
- Backup connections and tested failover
03Automation & Browser Systems
Repetitive operational work turned into supervised, recoverable pipelines.
- Browser automation built with Playwright
- Separate planner, executor and verifier
- Sessions and auth state held reliably
- Monitoring and recovery for long jobs
04Full-Stack Product Development
Making complex infrastructure operable by people who never read a log file.
- APIs, services and real-time interfaces
- Dashboards and operator control panels
- Auth, databases, queues and webhooks
- Deployed and maintained in production
05Infrastructure & Networking
The layer everything else quietly depends on.
- Linux, Docker and reverse proxies
- DNS, SSL, firewalls and monitoring
- GPU server and workstation planning
- Dual-WAN failover and resilient internet
06Product & Technical Ownership
Accountable for the outcome, not just the commit.
- Operational problems into technical systems
- Concept through to production
- Technology chosen on reliability and cost
- Production issues diagnosed under pressure