Description
Position at Resolver
The role of a Risk Detection AI Engineer is to maximise the automated classification of millions of pieces of content that come through our platform by creatively applying and integrating advanced Large Language Models (LLMs) and other Artificial Intelligence capabilities. This is a role central to everything that we do to protect children, brands, and platforms online. As a Risk Detection AI Engineer, you will utilise a deep understanding of the online risk landscape, as well as our risk detection capabilities, to assist in developing and optimising automated systems for identifying and classifying risks across various online platforms. Leveraging cutting-edge techniques for LLM prompting, fine-tuning for specific risk contexts, and combining AI outputs with other signals, you will work to enhance risk detection, discover new and creative applications of AI for classification, and support the design and continuous improvement of workflows and solutions to ensure efficient risk management and protection against many areas of online risk, prioritising innovative, automated AI deployment wherever possible. AI Integration and Optimisation;
* Understands how risk discovery is done and data is collected
* Collaborates with senior team members to integrate existing or emerging AI technologies (LLMs, GenAI, behavioural signals) into Risk Detection workflows
* Designs and implements effective LLM Prompts and AI-based classification criteria for risk detection
* Conducts data quality checks and feature usage analysis to ensure AI models are receiving optimal input
* Provides input on the technical requirements and expected output of AI models to upstream Machine Learning and Data Science teams
* Documents AI/ML component behaviour and performance for internal stakeholders
* Supports RD Config Engineers in the optimisation of workflows that incorporate AI models
* Ability to adapt, refine, and execute scripts (e.g., Python) for testing, analysis, and validation of AI component integration
Qualifications & Experience
* Bachelor's degree or equivalent experience
* Ability to input on the design and optimisation of solutions using advanced techniques like proprietary NLP, LLMs, GenAI, and behavioural signals
* Experience with the use and implementation of Generative AI solutions.
* Ability to write effective LLM Prompts for Risk Classification.
* Scripting experience (e.g., Python) is essential
* Query language (SQL) desirable
* Ability to work with high volumes of data
* Good understanding of core detection metrics (e.g., Precision, Recall, F1-score) and their business implications
* Proficient in compiling and interpreting confusion metrics
* Knowledge of Resolver's risk taxonomy
* Knowledge of how data is collected for each product and overall data flow
* Able to read and write JSON
* Keeping up to date with emerging tech
* Proficient in compiling and interpreting confusion metrics