Twenty years making complex systems work together. Now pointed at intelligence that's finally ready.
AI has only recently become good enough to do real expert work. Getting it to run reliably — without a person stuck in the loop — is a systems problem, and that's what LlamaPath is for: expert knowledge turned into operated systems clients can own or subscribe to. It comes from a pattern its founder saw over two decades across networking, cloud and enterprise technology: the breakthrough is real; making it dependable inside a business is the systems problem.
The breakthrough is real. The work is making it operational — reliable, governed and useful after the first successful run.
The old options don't fit experts. Hiring an engineering team is slow and expensive. Off‑the‑shelf software bends your work to someone else's tool. A project build can age from the day it ships. LlamaPath is a different shape: a platform that turns what only you know into an operated system — one you can own or subscribe to.
Mike Ciesla — Principal of LlamaPath.
Mike Ciesla started deep in one thing: networking — the discipline of how the layers of a system connect. From there his career became about orchestration. Years in enterprise technology taught him to marshal engineering teams, product specialists and executives toward one thing: getting the customer the outcome they actually needed. A different layer, same job — making specialist parts fit into a working whole, around the result that matters.
Cloud, then crypto, then AI — each a different shape, read with the same lens: how do the pieces connect, which patterns work, what's worth keeping. Today that instinct coordinates AI agents instead of human teams — pointed at the same place: your outcome. The tools move fast; the systems thinking is what lasts, and it's what LlamaPath is built on.
Where it comes from.
- UNSW
Engineering & research
Bachelor of Engineering (Electrical, First Class Honours) and Commerce. Published research on moving intelligence closer to where the work happens — presented at Stanford in 2009.
- 2005–2007 · Bulletproof Networks
Engineer — one of Australia's earliest cloud providers
Infrastructure, hosting, operations and security for live customer environments.
- 2011–2020 · Cisco Systems
Data centre & cloud — support, architecture, then commercial
From production troubleshooting to solution design to running enterprise relationships. Held a double CCIE (Routing & Switching, and Data Centre) — among the most demanding certifications in the field, and rare to hold both.
- 2020–2025 · Google Cloud
Enterprise AI adoption at scale
Finished as AI Sales Lead for the Global Cloud‑Native Enterprises segment, and twice named to Cloud Club — Google's President's Club. A close view of how large organisations actually put AI to work.
- 2026– · LlamaPath
Building the platform
The product of one observation: the challenge isn't building intelligence — it's building systems that can reliably operate it.
Four threads, one platform.
Research
Published work; a habit of moving intelligence closer to the work.
Infrastructure
Deep roots in networking and data centre — how the layers of a system connect and stay up.
Commercialisation
Bringing enterprise technology to market and into use.
AI adoption
A close view of how large organisations put AI to work in production.
A platform business built around owned assets.
LlamaPath develops and operates its own platform assets — the Assembly Engine and Agentic Operations — and commercialises them for clients. The platform does the work that would traditionally require a specialist engineering function, which is what lets us deliver a software asset and keep it running as a product, not a loose project.
We work with Australian small and mid‑sized businesses and deep domain experts — people with specialised knowledge and no engineering team — who want that knowledge turned into an operated system, with the right ownership or subscription model agreed up front.
Think your work could run as a system?
A short capability assessment is the way to find out.