Mr Mark Mcdonald1
1Google DeepMind, Australia
Biography:
Mark is an Engineer & Advocate in Google DeepMind's Gemini Developer Relations team. He works on making Google's ML software platforms and APIs a smooth experience for all developers, from new to advanced.
He has worked on a range of Google products, including the Gemini API, PaLM API, TensorFlow, Google Maps and even Santa Tracker.
Abstract:
The rapid pace of generative AI presents transformative opportunities for research, but practical skills in using these powerful tools are often a barrier. This workshop aims to equip researchers with hands-on experience using Google's latest generative AI models to enhance their research workflows, from text generation and multi-modal data analysis to building interactive AI agents.
This half-day interactive workshop will guide participants from no-code prototyping through to writing complex apps for deep research, processing unstructured data and automating custom tool.
Starting with the basics, attendees will use the Gemini Python SDK to progress through text generation, multi-turn chat, and advanced multi-modal capabilities including image, audio, video, and document processing. We'll also cover unstructured input, structured outputs, function calling with external APIs, native tools, and an introduction to the Model Context Protocol (MCP) for building agents using 3rd party tools.
By attending this workshop, attendees will get comfortable writing code with Google's Gemini 2.5 models for a wide range of tasks, including automated deep research, and will develop practical skills needed to use generative AI to their eResearch projects.
The workshop will use Python in a web browser (Colab/Jupyter notebooks), so it is suitable for anyone with a basic familiarity with Python.