KTP Associate in Multimodal Content Analytics, Machine Learning, and Rule-Based Generative AI

Candidate Profile: PhD in a relevant AI discipline, or minimally an MSc in AI related discipline with significant, demonstrable commercial or research experience in a related field.

Skills/ experience required include:

Robust knowledge in data science, machine learning (ML), and predictive analytics particularly in NLP
Experience using rule-based Generative AI (Gen-AI) models.
A proven background of working with large datasets, and exposure to commercial environments.
Knowledge of Retrieval-augmented generation (RAG) framework
Significant experience in solving language-based problems.
Industry experience alongside relevant academic background
Effective interpersonal skills across a range of stakeholders from engineering / technical staff to clients and non-experts
Desirable:

Knowledge of network analytics and voice analytics
User centric Graphical User interface (GUI) and User Experience
Producing academic publications of the highest standards
Technical writing and reporting
Skills in project management

Personal attributes:

An aptitude to learn and solve technical challenges.
The capability to work both independently and collaboratively.
Effective communication skills for transferring technology insights from Aston into D&G.
Strong skills in documentation, report writing and presenting for a range of audiences
Project management skills with the ability to develop workplans and work to deadlines

The project aims to implement the latest advances in Artificial Intelligence (AI) to develop a novel Intelligent Insurance Fraud Detection System (IIFDS). This will identify high-risk behaviours and patterns enabling the early detection of potentially fraudulent activity to transform Domestic and General (D&G)’s approach to insurance fraud.

Key areas of focus include:

Building risk-scoring models based on internal and external data to flag high risk individuals and/or claims early – potentially refusing sales or charging higher premiums
Network analytics: developing models to tie together individual and device details with the aim of identifying duplicate accounts (currently being used to mask fraudulent activity)
Call transcript analytics- to identify fraud indicators from customer phone calls and build Generative AI (Gen-AI) models to flag these indicators from transcripts.
Creating a fraud detection ecosystem which expedites the work of operational teams.
The project will apply the latest AI techniques in a transformative way for D&G, innovative to the industry, by mitigating risk and fraud in insurance claims and improving operational efficiency.