Research Fellow/Engineer - Generative models (IM)

Generative models, particularly diffusion models, are becoming ubiquitous and important in a number of fields. However there are many kinds of bias encoded in such models – some due to the choice of training materials, some are regional, and some are inadvertent.

In this project we detect the bias, measure it, and create technological solutions to alleviate or reduce it.

We will undertake cutting-edge research for generative and foundation models and release an open source toolkit which will have deep impact in this field.

Key Responsibilities:

• Participate in and manage the research project with Principal Investigator (PI), Co-PI and the research team members to ensure all project deliverables are met. Attend all meetings and report regularly.
• Take the initiative to ensure the project succeeds.
• Curate a dataset.
• Work with Co-PIs on the bias criteria (soft skills).
• Implement gAI diffusion model.
• Python coding for bias detection.
• Python coding for bias mitigation.
• Prepare and release open source toolkit.
• Contribute to public engagement.
• Assist in writing publications and maybe general articles.
• Assist in other duties consistent with the role of research fellow or senior research engineer.
• Occasionally (hopefully not often) assist with administrative duties (paperwork) related to the project.
• If required for your preferred future career, you may choose to undertake a small amount of teaching as part of your workplace training.

Job Requirements:

• Have a Masters or PhD and relevant experience from a recognised university.
• Have implemented, fine tuned or have in-depth experience with generative (e.g. GAN) or diffusion models.

Key Competencies:

• Skilled at Linux, Python and Machine Learning
• Possess strong analytical and critical thinking skills
• Show strong initiative and take ownership of work