Responsibilities
- Create advanced features powered by large language models, including risk identification, alert interpretation, conversation summaries, and AI-assisted review tools
- Apply model explainability methods such as SHAP, LIME, and attention visualization to ensure outputs are auditable, version-controlled, and repeatable
- Improve inference speed and reduce token usage to meet production efficiency requirements
- Build retrieval-augmented generation and large language model pipelines capable of analyzing web-scale datasets for risk insights
- Integrate legacy rule-based filtering systems using regular expressions to enhance model accuracy and performance
- Develop real-time data processing workflows using technologies like Kafka and batch processing with Spark
- Establish systems for tracking experiments, managing model versions, and automating deployment through CI/CD pipelines
- Monitor models for statistical drift, bias, and declining performance, triggering alerts when thresholds are breached
- Collaborate within multidisciplinary teams using agile development methodologies
- Partner with Product Managers, Site Reliability Engineers, and Compliance subject matter experts to enhance system reliability, adoption, and business impact
