Anjelo
Fernando.
I deploy and support networks for SME clients across Auckland. Most of my work sits where physical infrastructure, change control, and support reality meet: rack installs, switching, firewall changes, service validation, and documentation the next engineer can actually use.
Most of my work is on-site: rack installs, switching, VLANs, firewall changes, and service validation. Physical infrastructure has details that remote-only work misses.
I try to leave change notes and final-state docs where I can, because I've seen what undocumented environments cost the next person in the seat.
I prefer boring, maintainable setups over clever ones. A network that stays up is more valuable than a network that's interesting to look at.
Network Analyst at Netbridge Ltd, deploying and commissioning network infrastructure for SME clients across Auckland. Day-to-day that usually means on-site installs, config changes, validation, and handover.
I also handle the operational side: vendor quotes, procurement, asset tracking, support contracts, and license renewals. It's not the glamorous part, but it's where a lot of environments quietly fall apart without someone watching it.
What I work
on.
Change records, final-state diagrams, and handover notes are part of the job, not an afterthought. I've seen what happens when they're missing and the original engineer is unavailable.
Most jobs have a constraint worth understanding: why the DMZ lives here and not there, why a tunnel over an open port, why the simpler design over the clever one. I try to document those reasons alongside the config.
I follow jobs through: procurement, on-site install, config, testing, and handover. That last step is where a lot of work quietly gets lost, so I treat it as part of the job rather than an afterthought.
Things I’m building.
Problem: Wazuh was generating 30–50 alerts per shift. Reviewing them meant manually reading raw log lines, cross-referencing event IDs, and deciding what was worth escalating. First-pass triage was taking 45–60 minutes per batch.
Constraint: Client data couldn't leave the network. No cloud LLMs, no external API calls, no third-party summarisation services. Whatever ran the AI analysis had to run locally on hardware already in the lab.
Outcome: Triage dropped from 45–60 minutes to under 10 seconds in testing. Each alert batch now produces a plain-English summary with MITRE ATT&CK tags and a one-click PDF report for management review.
Migration & Deployment
Problem: The company's documentation was split across shared drives, email threads, and individually-named Word files with no consistent structure. Engineers were spending real time finding things that should have been immediately findable, and institutional knowledge was leaving with people when they left.
Constraint: Management required a SharePoint fallback—the platform couldn't be the only copy of the data. The server also had to be internal-only, which ruled out any public SaaS wiki.
Outcome: Hundreds of legacy docs migrated into a single searchable BookStack instance with a consistent shelf/book/chapter hierarchy. The SharePoint export pipeline runs nightly, so there's a readable backup even if the server goes down.
Problem: I wanted a place to write about work without handing the content to a third-party platform I don't control. Shared hosting felt like a step backwards given the infrastructure I was already running at home.
Key decision: Publishing through a Cloudflare Tunnel rather than opening inbound firewall ports. The home IP stays hidden, there are no open ports on the router, and I don't have to deal with dynamic IP or residential ISP restrictions.
Outcome: A hand-coded static site and a Ghost blog, both served from a mini PC at home, with zero external hosting dependencies and full control over the deployment pipeline.
Containerized Infrastructure
A working lab for running real services, testing configurations before production, and learning what breaks under load. Not a demo environment—things actually depend on this staying up.
All services run in Docker Compose with version-controlled configs, meaning the full environment can be rebuilt from scratch in under an hour. That constraint forces clean separation between state and configuration.
Qualifications.
Let’s connect.
Open to connecting with people in networking, infrastructure, and IT operations. If you want to connect and talk shop, get in touch.