OUR
SPEAKER
Thushan Fernando
Senior Consultant
Thushan is a Senior Consultant and AI practitioner with over 20 years of experience building software and exploring emerging technologies. He speaks regularly on practical AI engineering, from Claude Code and agentic development through to fine-tuning models, running LLMs locally and at the edge, optimising CUDA workloads and serving models at scale with technologies such as vLLM and SGLang.
He has spent more than a decade working with machine learning and has hands-on experience improving model and token efficiency and helping customers find practical, cost-effective ways to use LLMs. Beyond AI, Thushan has presented on Go, TinyGo, Rust, containers and systems development. His talks focus on understanding how technology works in practice, cutting through the hype and sharing lessons from building and experimenting with it firsthand.
TALKS
Effectively Using Claude Code
This session cuts through the surface-level hype to give developers a practical, working understanding of Claude Code: what it actually is, how it thinks, and how to use it well. We start with the mental model: why Claude Code is fundamentally different from autocomplete tools, and why that distinction changes how you should work with it. From there, we cover the mechanics that matter: context windows, token usage, and why getting these wrong quietly undermines your results.
The bulk of the session is hands-on and workflow-focused: plan mode, CLAUDE.md, sessions, and slash commands as a daily driver; then extending Claude Code with skills, custom commands, subagents, hooks, and MCP servers. Thushan will be honest about what works and what doesn't, the patterns that unlock real productivity, and the anti-patterns that tend to bite. The session closes with a set of concrete recipes you can try right away, so attendees leave with something immediately actionable, not just inspiration.
