Back in 2020, 15 years into his career, then-platform engineer Sam McLeod was finding his work world, in his words, “pretty boring.” But instead of settling, he leaned into his curiosity and started diving deep into the world of AI, allowing him to take control of his career and the wave of change that was about to impact his industry tenfold. Now, he’s a leading voice in AI for engineers at Mantel, and his journey proves that you can transform your professional life by just having a go.
Sam’s path to AI wasn’t a straight line. He started in tech right out of high school, doing computer repairs and server engineering in New Zealand before moving to Australia and heading up platform engineering at a not-for-profit. He was an early advocate for DevOps, a cultural movement that brought developers and operations teams together. However, by 2014, the once-exciting field of platform engineering had begun to feel stagnant.
That’s when he started playing around with open-weight AI models, and early in its limited beta, he got hands-on with GitHub Copilot. This was a turning point. He saw firsthand how AI could accelerate and automate the tedious parts of coding, which resonated with his long-held passion for automation and optimisation. He was hooked.
Sparking Curiosity: A DIY Approach to AI
Sam’s journey shows that you don’t need a formal education in machine learning or a specific title to get involved with AI. It starts with curiosity and a willingness to get your hands dirty. He found that the most effective way to learn was through self-directed learning.
“I definitely learned by breaking things and then figuring out how to put them back together,” he says. This hands-on approach is at the core of his recommendations for anyone looking to get into AI.
Recommendations for Getting Started with AI:
- Dive into Open Source: The open-source community is a goldmine for learning. Find AI-for-engineering projects that interest you and contribute. Even small pull requests to fix a bug or add a new feature can provide invaluable experience.
- Join Communities: Communities like the Reddit forum “Local Llama” are dedicated to running large language models on your own hardware. Being an active member in these groups exposes you to the latest tools, methods, and models.
- Build Your Own Home Lab: Sam set up a home server to run his own AI models, which allowed him to learn about the low-level technical details. While you don’t need to go this far, having a personal space to experiment is crucial.
- Focus on the Fundamentals: Rather than getting bogged down in specific tools, understand the basics of how large language models fundamentally work. Think of them as prediction engines. Every piece of information you give it adds to its ability to make the next prediction. When you understand this, you’ll be able to write better prompts and get better results from any AI tool.
- Just Have a Go: “Rather than getting bogged down in blog posts and articles – grab the latest agentic coding tool (he recommends Cline or Claude Code) and have a go at creating a basic hello world app.” Sam says. The best way to learn is by doing. Start with something small and simple, and don’t be afraid to fail.
AI in Action: Real-World Impact
Sam’s advocacy and hands-on work have had a significant impact at Mantel. He’s moved from advocating for AI from the sidelines to being the “AI guy,” leading technical pre-sales, agentic coding training and R&D for clients. He’s seen firsthand how AI can transform projects and unlock new capabilities.
For one client back in mid-2024, he built an agentic system to reverse-engineer a large, 23-year-old, undocumented legacy codebase. What would have taken months to understand was made clear in just a few days. In another instance, he showed a team how to redevelop and modernise an application in a matter of weeks, leveraging agentic coding and tools like Cline and MCP, even with an engineer who had no prior experience with the programming language.
“That’s probably the first project I was on where there was a realised value beyond basic copilot style acceleration,” he says, “showing the client what’s possible.”
These stories aren’t just about speed; they’re about empowering people and teams to do what they thought was impossible.
”I was an early tester on a bunch of AI tools and services, and I thought: "This is a great opportunity for me and Mantel." I shared that with the company, they said "That's interesting, tell us more..." and here we are today! The way my individual journey has been supported by Mantel has been really valuable to me.
Sam McLeod | MantelAI Engineering Principal & CSP
The Future of Work and Learning
The traditional career path is changing. With AI automating many jobs, it’s more important than ever to be adaptable. Sam believes that the key to staying ahead is maintaining that sense of curiosity that drives you to learn and explore new things.
His advice to anyone feeling stuck is simple: spark your own curiosity. Find something that excites you and gives you energy rather than taking it away, even if it’s outside of your current role. When you’re genuinely interested in something, learning becomes a passion project, not a chore. That’s when your learning accelerates, and you open the door to new opportunities.
If you’re looking to get started, check out Sam’s page here for some more insight.
AI at Mantel:
Today, Sam and Mantel’s growing team of AI specialists, support clients from strategic consulting to hands-on implementation and training. In addition to helping clients with the complexities of AI, Sam and the team have been pivotal in educating those around them.
Internally, they ran an AI Enablement Series: to ensure our people understand what AI means for Mantel and our clients, and can spot the new opportunities this technology unlocks. A range of AI learning streams were offered to suit different capabilities from Data Scientists to specialised engineering roles.
Recognising that everyone learns differently, they’ve built AI skills through a variety of avenues, from targeted 1:1 learning and Vibe with Me sessions (where they demonstrate real-time AI solutions for practical use cases), to 20+ brownbag sessions with thousands of attendees and a hands-on testing environment, which saw 223 active users in the first month alone.
