AI FOR ENGINEERS

“AI For Engineers: Stop Piloting. Start Scaling. Master AI for Real Engineering Impact.””

Equip engineers with a critical, technically-grounded understanding of LLMs, Generative AI, and Agentic AI—cutting through the hype to reveal what these tools actually do, where they excel, and where they fall short.

OUTCOMES:
By the end of this half-day course, you will be able to:
• Construct effective prompts that get reliable results from common LLM tools
• Make informed decisions about what data to share (and what to keep private)
• Apply Claude to a real-world data analysis problem
• Critically evaluate and verify AI-generated output

Part 1 – What is AI? What it Isn’t

• No intelligence here
• Training vs Inference
• Ethics of Training
• Generative AI, Agentic AI, LLMs
• Current Limitations
• Just a Tool that’s Good at the Middle Bit
◦ No original thought
◦ Not great at the final polish
◦ Always verify
• Context and Training
• Tokens
• AI in the Engagement Economy
◦ AI is not your friend
◦ AI wants you to keep putting tokens in the slot
◦ The AI ego massage

Part 2 – Which Tool?

• Model Selection
◦ Claude, ChatGPT, Gemini, Copilot
◦ Chinese Models – Deepseek, Qwen
◦ Running your own models – Llama
◦ Comparative test
• Everything Else
◦ Embedded tools
◦ Specialised tools: DeepL, Gamma, Leo

Part 3 – Safety and Security

• What gets shared, what gets retained – do we believe the companies?
• What to share, what not to share
• On-prem vs off-prem
• When to go local
• Copyright and derivative works

Part 4 – Getting the Best from Claude

• Why we are using Claude
• Chat, Projects, Code and Cowork
• Opus, Sonnet and Haiku
• How to construct a prompt
◦ Product Manager/Toddler Parent
◦ Markdown formatting
• Copyediting example
• Skills, connectors, extensions and MCP
• How to know if your Claude is going off the rails
• The confidence problem
◦ You have insight, Claude just has (out of date) knowledge
◦ Don’t let it convince you when you know better
◦ It is spectacularly bad at estimating timescales
• Using Claude to set up Claude
• Using Claude to verify Claude

Part 5 – Practical Data Analysis with Claude

Problem – Data Centre Impact: You are conducting a feasibility study on building a [X MW] data centre in [randomly chosen community]. You need to quickly assess:
• Energy and water consumption of the project
• Cooling options
• Solar and battery requirements to power it
• Estimated operating costs

You have 30 minutes to crunch the numbers and produce a presentation.

Part 6 – Review Output

• Review student output
• Questions and final wrap

You’ll learn from an academic lens with a tried and tested hands-on concise analysis and real-life challenges in implementing AI into your workplace ethically and with Truth.

TRAINER:

Ed Lynch-Bell:

Ed Lynch-Bell doesn’t teach theory; he teaches the brutal realities of scaling clean energy technology in a world where 90% of pilots never reach commercial deployment. Over 15 years, Ed has taken innovations from the laboratory to large-scale manufacturing (Aquion Energy’s sodium-ion batteries), launched products that defined markets (AGL Energy’s home
battery program), established safety standards that became national regulations (Australia’s lithium-ion battery protocols), and built infrastructure networks at a commercial scale (Evie Networks’ DC fast-charging rollout—Australia’s largest).
Engineering Educated at Carnegie Mellon USA, Politecnico di Torino, Italy, and University of Sheffield UK, Ed now runs Second Mouse, a boutique consultancy that helps corporates and startups navigate the gap between “promising technology” and “bankable business”. Ed is the co-founder of EV MEETUPS events organised in Melbourne, Adelaide, and Sydney for the Electric Vehicle industry and government policy.

What makes Ed’s training different:

He’s made every mistake founders make, survived every investor grilling, and knows exactly what separates funded projects from PowerPoint dreams. Plus, he is a hands-on, test-and-tried AI expert in
business management.