Kyle Dunkerley
Open to roles

In Melbourne.

Connect or message

Infrastructure leader · Melbourne

Kyle Dunkerley

Kyle
Dunkerley

I build capable teams and dependable infrastructure.

Associate Manager, Infrastructure Services · DXC Technology

271 servers · 4 engineers · 100+ changes run · DR timed every six months, none failed

Path · scope grows with every step
  1. 2007LAN AdministratorKwaZulu-Natal, South Africa
  2. May 2011Helpdesk to on-site supportProphecy Networks, NZ
  3. 2018Microsoft Solutions SpecialistProphecy Networks, NZ
  4. Aug 2019Systems EngineerLANtech / Spark, NZ
  5. 2021 to 2026WinTel Admin → Team LeadDXC Technology, Melbourne
  6. 2026 → nowAssociate ManagerDXC Technology
Kyle Dunkerley
What I’m looking for

People. Platform. Direction.

An internal infrastructure or IT operations leadership role in Melbourne.

Location

Melbourne

Hybrid or in the office. Melbourne CBD, Docklands or the eastern suburbs. Full work rights in Australia.

The work

An estate with room to improve.

Set a practical roadmap, make operations dependable and help engineers take ownership.

Experience

Leading since 2021

People development alongside delivery and escalations.

Longer term

Broader technology leadership

Grow into senior IT management and, eventually, CIO responsibility.

What the team says

“Even routine or seemingly boring tasks often turn into valuable challenges that push us to learn, grow, and improve our skills.”

A member of my team

Start a conversation

Tell me about your team.

Four weeks’ notice once an offer is accepted. Connect or message on LinkedIn.

Kyle Dunkerley
Hybrid estate

Dependable today. Easier to run tomorrow.

Server operations, senior escalations and lifecycle work across Australia and New Zealand.

Estate

271

Servers: physical, on-premises and Azure.

Employees in the business

5,000+

Across Australia and New Zealand.

Operating rhythm

Change. Recover. Learn.

  • CAB-reviewed changes
  • 100+ changes raised and run through CAB
  • DR test every six months, none failed
  • Written incident reviews
Capacity decision

4 → 2 vCPUs

Right-sized three batch-processing VMs in Azure from 4 to 2 vCPUs when the workload dropped, which cut the cost.

Next improvements

Less routine upkeep

Moving a 3 TB file and print server used by 150 to 200 admin staff to Azure, with DFS namespaces to bring file sharing up to standard, and rebuilding the production information server for a paper mill on Windows Server 2025. Exploring Server Core and modern app delivery.

Hardening

Harden by default.

Switched off SMBv1 and IE11, retired TLS 1.0 and 1.1 with TLS 1.3 as the standard, and turned on LSA protection, all through Group Policy.

Production

A platform a mill runs on.

I own patching for the 18 production servers behind a paper mill, from the vendor’s approved update list and a patch calendar agreed with the site before each shutdown. In 2026 I built new terminal and web servers and moved the session broker off the database server before the June shutdown.

Delivery at scale

12 sites, about 35 cutovers.

A server refresh across 12 sites in Australia and New Zealand, then about 35 cutovers consolidating site file servers into central and Azure storage.

Kyle Dunkerley
Team

Build a team that owns the work.

Clear priorities, practical mentoring and trust earned through follow-through.

One team

4

Two in Victoria and two in Manila.

My role

Set direction

Prioritise, take the hardest escalations, run meetings when my manager is away.

A real progression

Service desk → ownership

  • Joined without server experience
  • About a year working side by side
  • Now owns patching end to end
How we work

Visible priorities. Living notes.

Fortnightly sprints, shared Kanban, runbooks and documentation the team can update.

What the team says

“No matter how busy things get, he takes the time to listen, share his knowledge, and help others succeed.”

Ray, DXC

“Kyle is a professional, supportive, and dependable colleague and team lead who is always willing to help the team.”

Marvin, DXC

Kyle Dunkerley
AI & toolkit

Give engineers more reach.

I’m pro AI where it helps people investigate faster, build better and take on harder work.

Working rule

Draft. Understand. Test.

The engineer challenges the answer and owns the result.

Identity & compute

The foundations

Automation & visibility

Tools with a purpose

Platform foundations

Enable the next step.

Kubernetes test environment for SAP. Power Platform production handover.

Kyle Dunkerley
Credentials

Learning that comes back to the team.

Azure architecture and administration, built on a Windows Server foundation.

Azure Solutions Architect Expert

AZ-305

Issued October 2024, renewed through October 2027.

Azure Administrator Associate

AZ-104

Issued September 2024, renewed through September 2027.

Fundamentals

AZ-900

June 2024.

Always learning

Certified since 2009

Seventeen years of Microsoft certification, from Windows Server to Azure.

Education

IT Networking

Two-year diploma, Varsity College. Completed in 2006.

Kyle Dunkerley
Selected work

Decisions that changed the work.

Choose a story: the trade-off, my responsibility and what happened next.

Kyle Dunkerley
How I think and lead

Build capability. Reduce dependence.

Better infrastructure leaves the team more capable of running it.

Responsibility

The buck stops here.

I own the direction, back the team and help put things right.

What the team says

“He leads by example, maintains a positive attitude, and is always ready to lend a helping hand.”

A member of my team

Communication

Call when it matters.

Make space for questions when a problem is urgent or tangled.

Trade-offs

Make the case. Accept the call.

An app vendor would not test on Server 2025, so two servers stayed on 2022. I made the case, accepted the call and moved the rest to 2025.

Decision

Hold the patch.

When a Microsoft monthly update broke remote desktop on our test servers in September 2026, I recommended rejecting it. A production planning system stayed up and the monthly reboot cycle carried on.

Decision

Recover what matters.

During a file server migration to Azure I found about 62,000 corrupt files and recovered about 58,000 of them (250 of 260 GB) from clean restore points. The business signed off on the rest, which were untouched and covered by older backups.

Kyle Dunkerley
Beyond the day job

Different interests. The same care.

Community, conversations, software and music. Each worth doing properly.

The XboxCastSince 2016

Make room for conversation.

A gaming and technology podcast. I rebuilt its website off WordPress.

Southlight I/O2026

Build the idea.

Apps that work offline, need no account and send nothing to me, because I run no servers. Founded in 2026 and built with AI-assisted development.

Board service2023 to 2025

Southern Cross Church

Shared decisions on finances, remuneration, equipment and events.

Writing

The personal side

Essays and projects.

Kyle Dunkerley
Made By Humans

Music with energy. Room to create.

My independent instrumental rock and metal project, separate from Southlight I/O.

Five-track EP

InstruMETAL

Hope, tension, struggle and victory. An energetic alternative to lo-fi.

Listen

Apple Music

Made By Humans artist page.

Listen

Spotify

Made By Humans artist page.

The sound

Heavy meets cinematic.

  • Guitars and drums
  • Electronic textures
  • Space to work or create
First single

Victory Spurs Us Onwards

The feeling of a hard-won finish.

The story & creator use

Why I made it

The introduction and intended use for creative work.

Kyle Dunkerley
Windows Insider

Testing what was coming next.

Two years as a Microsoft MVP, and a community I helped build in New Zealand.

Microsoft MVP

2016 to 2018

Stepped back in 2018 to focus on my own goals. Kept testing as an Insider.

Top Insider, New Zealand

5,000+ upvotes

On my feedback. Named New Zealand’s Top Windows Insider in 2016. I tested builds and wrote up what broke.

Community

Meetups and Ignite

Hosted Insider meetups, met Microsoft staff and went to Ignite several times in New Zealand. Named on the official Windows Insider blog.

Channel 9 · 2016

On Microsoft’s Channel 9

Interviewed at Ignite New Zealand in 2016 alongside Dona Sarkar and Jeremiah Marble. Also pre-release calls with Microsoft staff.

SeniorNet · 2017 to 2019

Adviser on Saturdays

Talks on what was changing in Windows 10, plus tech support in nursing homes. Mostly helping people stop buying antivirus every time a pop-up asked.

At work

Lunch and learns

Showed colleagues what was coming to Windows 10 before it arrived.

Kyle Dunkerley
Career & perspective

Choosing the difficult move.

South Africa to New Zealand to Melbourne. Starting again and taking on more responsibility.

2007 · South Africa

LAN Administrator

Hands-on support for 120 lab PCs.

May 2011 · New Zealand

Prophecy Networks

Started as a Helpdesk Administrator on a 12-month fixed-term contract. Promoted to Helpdesk Engineer on 1 June 2012, then moved on site at Napier Port.

September 2013 to October 2018 · Napier Port

One role. Growing scope.

A 24-hour port with about 300 staff. Led the Windows 8.1 and Office 2013 rollout across 250 machines, moved the port from on-premises Exchange to Office 365 and became main admin for its Office 365, Exchange Online and Azure tenant. A say in hardware purchasing. Scripted the overnight recovery of about 60 deleted accounts. No new title.

2018 to 2019 · Prophecy

Microsoft Solutions Specialist

Operational experience informing identity and access solutions.

August 2019 · LANtech

Earn the client’s trust.

Systems Engineer, later Senior, with no formal promotion. Statements of Work.

2021 → now · Melbourne

Engineer → Lead → Manager

Contracted to DXC via Ploy from January 2021 for a business separation. WinTel Team Lead from November 2021, then Associate Manager, Infrastructure Services in June 2026. DXC’s HR title for the lead role was Professional System Engineer; the client-facing title is still WinTel Team Lead.

Kyle Dunkerley
PIR-01 · Team capability

Rebuild around the engineers.

Staffing changes brought Windows infrastructure into a wider end-user computing group.

The team today

4 engineers

On-site and offshore delivery.

The outcome

Ownership grows.

No server background to running patching end to end, in about a year.

Starting point

Build a working team.

Initially one on-site and two offshore engineers, with server capability to develop.

What I introduced

A shared direction

Sprints, Kanban, living documentation and mentoring in Azure and IaC.

The hard part

Earn trust through delivery.

A methodical roadmap and follow-through mattered more than my title.

What the team says

“One of the most dependable and professional colleagues I have had the privilege of working with.”

A member of my team

Kyle Dunkerley
PIR-02 · Business separation

Move the estate. Keep the business moving.

New infrastructure, a new domain and a deadline across Australia and New Zealand.

Physical hosts moved

50+

Virtual machines moved

~100

My remit

Delivery end to end

Standards, installation, outage windows, cutovers, decommissioning. Not procurement.

Build

Agree the standards.

Prepare the new estate while the old stays in use.

Move

Join up the timetable.

Coordinate stakeholders and planned outage windows.

Finish

Retire what we replaced.

Complete operational cleanup as part of the delivery.

Kyle Dunkerley
PIR-03 · Automation

Readable is maintainable.

Automation improves an estate when more than one person can safely change it.

The change

ARM → Bicep

Reusable deployments, with backup and DR modules.

The result

Knowledge beyond the author.

Consistent deployments and a clearer route for junior engineers to maintain them.

Before

One person’s memory

Complex scripts concentrated knowledge and made changes harder.

After

Shared understanding

Simpler PowerShell, documentation and teaching.

The design test

Can the next engineer support it?

Readability is part of the engineering.

Kyle Dunkerley
PIR-04 · Incident recovery

Restored. Still locked out.

A mission critical Server 2012 VM in Azure, past end of support, came back from backup but could no longer authenticate.

Service restored

~3am

Recovery completed after a long incident.

The lesson

The VM is only one part.

Restoring authentication was what returned the service to use.

Investigate

Work through the trust.

The restore went back far enough that LAPS and admin passwords had changed, and the domain trust was broken.

Recover

Reset the computer account.

Used the Azure run command and PowerShell to reset it and rejoin the domain.

Communicate

Stay on the bridge.

Keep stakeholders informed.

Follow through

Record the cause.

Documented the root cause and recovery in the incident report.

Kyle Dunkerley
PIR-05 · Migration · April to June 2025

When the cutover changed.

I led planning, cutovers and health checks for a Hyper-V to VMware move: a three-week cutover in a project running April to June 2025.

Scope

21 VMs

Five old Dell Hyper-V hosts (three production, two DR) to four Lenovo VMware hosts (two and two).

SQL server

3.5 TB

Total server storage across five drives.

The interruption

~6 hours on paper

The conversion locked the database at about 20%. Production stopped.

Option 1 · Stop

An uncertain release

Lose scan progress with no guarantee the database lock releases.

Option 2 · DR

Reconcile later

New orders would need merging. The DBA identified a risk of missed orders.

Option 3 · Chosen together

Complete the migration.

With the site manager and DBA: adjust, continue, finish. About four hours saved.

  • Next: rebuilt two smaller database servers rather than converting them.
Kyle Dunkerley
PIR-06 · Practical AI

AI finds a pattern. Engineers make the call.

A useful starting point for an investigation, backed by people who can challenge the answer.

Copilot analysis

1,081 reports

Backup reports reviewed for repeated failures.

My contribution

Set up the investigation.

Asked an engineer to verify the findings and the cause.

Team delivery

A justified change

Copilot found a job failing at the same time on the same day, and servers that had never completed a full backup. Moving the schedule and fixing the gaps repaired three of about ten backup jobs; later reports showed successful backups.

A separate patch review

Some findings did not hold up.

The team found superseded updates and software that wasn’t installed.

What the team says

“He encourages us to share ideas and explore new technologies that help resolve issues more efficiently.”

A member of my team

The responsibility

Understand. Test. Own it.

A confident generated answer is the beginning of engineering work.

Kyle Dunkerley
PIR-07 · People

Take myself out of the queue.

Everything used to wait on me. Now the team sends me only what it cannot solve.

Incidents I handled a year

~⅓ less

The team solves more on its own, and I took on the projects and upgrade work my manager handed over.

Starting point

Everything waited on me.

About 436 incidents a year, and every one of them came through me.

What I did

Only what they cannot solve.

The team sends me the Level 3 and 4 tickets they cannot fix. Everything else is theirs.

The hard part

Write it down.

Letting go of the queue meant documenting every fix, so the next similar ticket does not wait on me.

How I think about it

Make myself unnecessary for the routine work.

So I can take on the bigger stuff.

Kyle Dunkerley
PIR-08 · Diagnosis

Everything worked except the join.

New deny-by-default firewalls went in around the site server networks, and Windows domain services started failing in ways that looked contradictory.

The puzzle

It looked fine.

The domain join still failed, late in the process.

The two missing ports

445 / 135

SMB and RPC.

What I did

Narrow it down.

I traced it to the SMB and RPC stage and asked the network team to check both ports. They were missing, and they were opened.

The hard part

Keep testing.

Opening the ports was only the first step. I worked through each dependency until the site servers were fully managed again, and the team documented the full set of Windows ports a site server needs.

The lesson

Domain join is more than DNS and LDAP.

Most of my job is problems that sit between Windows, network and security teams.