EPISODE 02

TIKS Solutions → Damstra Technology

Desktop Support Engineer · Aug 2021 – 2024

Joined a startup as Desktop Support. Left having engineered production systems, led a team, and shipped software to 300+ devices.

Better to ask for forgiveness than permission.

A startup. One IT person. 300 kiosks. 100 Raspberry Pis.

TIKS Solutions was a startup building workforce management kiosks. When I joined, the infrastructure was held together with manual processes, TeamViewer remote access, and configuration drift. I was the only IT person. That's not a complaint — that's the opportunity.

What I owned

  • Fleet management: 300 Windows kiosks
  • Hardware automation: 100 Raspberry Pi boom gates with relay control
  • Remote access: VPN + SSH → evolved to SaltStack
  • Device management: TeamViewer → ManageEngine Endpoint Central
  • OS deployment: Manual → master image workflow
  • Support: L1 + L2 + L3 (I was all three)

Problems solved / Key contributions

Disk Space Crisis

Root Cause, Not Clear Tickets

120GB kiosks partitioned wrong on imaging. Tickets flooded in. Traced the pattern, found root cause, extended volumes via Diskpart across the full fleet. Fixed the imaging process. Zero recurrence.

Situation

Windows kiosks in the field were imaged with a 120GB OS partition that was too small for real workloads; logs and apps filled disks and the support queue spiked with recurring incidents.

Task

Stop the immediate pain across the fleet, prove root cause (not one-off hardware), and change the imaging pipeline so new deployments never shipped with the same defect.

Action

Correlated tickets to a common partition layout, used Diskpart safely to extend volumes on affected devices, then corrected the master imaging workflow and documented the fix for ops.

Result

Fleet stabilized, the incident class stopped repeating, and new builds inherited the corrected layout — tickets dropped from a flood to zero for that category.

Surface Pro Dead Device Tickets

Eliminate the Category

Kiosk enclosures had no auto power-on. Users thought devices were dead. Researched UEFI-level capabilities via Surface Configuration Tool. Enabled auto power-on. Entire ticket category eliminated.

Situation

Surface Pros in sealed kiosk enclosures looked "dead" after power events because hardware did not auto-resume; operators opened tickets for hardware replacement before checking UEFI behaviour.

Task

Remove false-positive hardware churn and give field staff confidence that devices would come back online predictably after maintenance or power cycles.

Action

Researched Surface UEFI capabilities, used Surface Configuration Tool to enable auto power-on where supported, and validated on representative kiosks before rolling guidance fleet-wide.

Result

The entire "dead device" ticket category disappeared — faster recovery in the field and less noise for support and logistics.

No IVR Phone System

No System? Build One.

Company had no phone routing. Researched, proposed, and built AWS Connect + Lex bot from scratch — self-directed, no prior experience. Professional IVR deployed at startup cost.

Situation

A growing customer base was calling in, but the company had no professional call routing — only ad-hoc forwarding — which hurt credibility and made handoffs messy.

Task

Deliver a credible IVR with sensible routing and self-service entry points without committing to heavy proprietary telephony spend.

Action

Self-studied AWS Connect and Lex, proposed a lean architecture, implemented queues and a Lex-driven front door, and wired it into how support actually answered the phone.

Result

A production IVR at startup-friendly cost: callers got a professional first impression and the team stopped improvising on every ring.

Electron CI/CD Pipeline

Commit to Deploy in Minutes

Built a full Bitbucket Pipelines CI/CD workflow for the Electron app. On every commit: pipeline triggers, builds the production Windows EXE, pushes to cloud storage, sends a Teams notification with the download link. Zero manual release steps.

Situation

The Electron kiosk client shipped through manual builds on workstations — easy to skip steps, ship the wrong binary, or leave the field waiting with no audit trail.

Task

Make every merge produce a single trusted Windows artifact, published in a known place, with the team notified automatically.

Action

Implemented Bitbucket Pipelines: build the production EXE on each qualifying commit, push the artifact to cloud storage, and post a Teams message with the download link for installers.

Result

Commit-to-installer in minutes with zero manual packaging for routine releases — fewer mistakes and predictable field updates.

Flagship project

★ Electron Kiosk Desktop Application

Production Engineering — Solo · Flagship

Chrome flags broke with every update across 300+ customer-facing kiosks. I identified the problem, built a configurable Electron app with CouchDB remote config and Teams proactive alerts. Approved by Solution Architect. Deployed to full fleet. Then built a complete CI/CD pipeline on Bitbucket.

ElectronCouchDBJavaScriptMS Teams APIBitbucket Pipelines

Situation

300+ customer-facing kiosks depended on Chrome-style unattended modes and flags that broke silently on vendor updates — outside our control but inside our SLA.

Task

Replace a fragile browser-shell approach with a maintainable kiosk runtime, remote configuration, and operator visibility that could pass architecture review.

Action

Built a configurable Electron client with CouchDB-backed remote config and proactive Teams alerts; secured solution-architect sign-off; rolled out fleet-wide; then automated builds and releases on Bitbucket Pipelines.

Result

A stable kiosk surface ops could steer without waiting on browser vendors — faster iteration, fewer Sev-2 surprises, and a repeatable release path.

Skills this episode

Electron
Strong
ManageEngine
Strong
Grafana
Strong
Golang
Proficient
Docker
Proficient
SaltStack
Proficient
PostgreSQL
Proficient
AWS Connect + Lex
Proficient
Bitbucket Pipelines
Proficient
GitHub
Strong

Grafana, Zabbix, Sentry; Golang network monitor with Slack alerts; SaltStack for RPi fleet.

Next episode

IdeaGen