Unravelling Machine Learning: Technical Challenges and Ethical Minefields

For our August meetup, we will be welcoming two fantastic speakers to present on the topic of automated systems from two different perspectives: the first a technical deep dive into the world of data engineering, using a case study from the real estate sector; and the second a look into the ethical and human-centric complexities of data science solutions.

------
Talk #1: Delivering personalised property recommendations in real-time

Presented by Kamal Hossain
Developing a real-time personalisation system for property listing recommendations is challenging. The focus extends beyond AI/ML model training to encompass data engineering and data integration. The dynamic nature of property listings (new properties being listed and properties being sold), combined with evolving user behaviour while browsing through properties; demands real-time personalisation of listings.

To address these challenges, the engineering team needs to overcome multiple obstacles. In this session, Kamal will share the characteristics of recommendation systems driven by machine learning, delve into the associated engineering challenges and how to deliver a successful end-to-end implementation.

Speaker Bio:
Kamal Hossain has more than nine years of experience in cloud based scalable end-to-end AI/ML system design, AI/ML model development, data platform and data engineering. Currently, he is working with CMD (Mantel Group) as a Lead Data Consultant. Previously, Kamal was a Machine Learning Engineer with BHP and the Lead of Data Science and AI with PropertyGuru Group in Singapore.

------
Talk #2: Designing Responsible Human-Centric Data Science Solutions

Presented by David Smith

Automated systems have caused vast amounts of harm around the world, and an Australian example of this is the "Robodebt" scheme, which has been described as a "shameful chapter" and a "massive failure in public administration" leading to $1.2b of compensation and a Royal Commission.

How do we stop this from happening again?

This talk will focus on the principles and practical steps to implementing human-centric, responsible data science solutions and explore the concepts relating to algorithmic decision-making, with a focus on fostering responsible and ethical implementations.

David will highlight the importance of human judgement and accountability in algorithmic processes, urging caution when deploying fully automated systems, delving into the principles of fairness, dignity, and justice, with a focus on equitable outcomes for vulnerable groups.

Speaker Bio:
David Smith is a passionate technology professional with a diverse background in architectural and computational design, software development, and data science. Holding a Master of both Architecture and Data Science from UWA, David has a proven track record of leveraging his unique skill set to solve complex problems, drive innovation, and bridge industry silos to generate value for companies.

Currently working as the Programs and Digital Executive at the Centre for Entrepreneurial Research and Innovation, a non-profit organisation focused on commercialising deep technology research, David utilises his expertise in digital technologies, innovation, and communication to empower industries of the future.