Dr Gnana Bharathy1, Ms Kathryn Unsworth2
1Australian Research Data Commons (ARDC), Sydney, Australia, 2Australian Research Data Commons (ARDC), Melbourne, Australia
Designing an AI & Analytics skills training pathway is a challenging issue, owing to not just the diversity of students and subspecialties, but also the need to bring in a range of professionals at different stages of the research project.
In this context, we catalogued and assessed the existing skills training courses in this space. We did this according to the stages in the data science or analytic project lifecycle, and determined the key issues and skills required at these stages, including problem formulation, data acquisition, data pre-processing and exploration, model development and validation, developing the proof of concept, and finally iterating further to production. We elicited best practices such as: need for facilitating stakeholder participation in an empathetic manner, and understanding the fragile nature of the path to production to enable translation and impact. Such practices should get translated into skills development too.
We integrated such best practices into a draft framework, which is currently being validated. This framework enables us to identify and assess the training tools and activities, visualise the gaps, and manage the distribution of training activities.
This work-in-progress exercise would develop inputs for curriculum design for AI & analytic roles, and would define key capabilities that are required for an effective analytics team, including domain expertise, business analysis, architecture, data engineering, statistics, data analysis, modelling, communication and team leadership. This is a deviation from the popular assumption that analytic roles such as data scientists are unicorns, who could effectively perform all these functions.
Biography:
Gnana Bharathy: Gnana works as AI/ML Research Data Specialist at the Australian Research Data Commons (ARDC) and is a practitioner, researcher and mentor, in AI and advanced analytics space. Gnana has been a Data Science and Design Consultant in the industry for over 15 years, has served a range of clients, and has conceptualized, developed and deployed several products, in health, energy, manufacturing, supply chain, environment, political and government sectors. Gnana is also engaged in application and practice-based research, including in the AI risk and lifecycle issues. ORCID: https://orcid.org/0000-0001-8384-9509
Kathryn Unsworth: A librarian in a previous life, Kathryn has transitioned from information to data to data skills. She works to ensure the ARDC’s leadership in the skills space provides a cohesive national digital research skills agenda for the Australian research sector.