GovTech Anti-Scam Products (GASP) is a product and engineering group that builds tech products to disrupt and deter scammers at scale.
We are building advanced analytics, deterministic rules, and machine learning models to detect, disrupt, and prevent fraudulent activity in the scam ecosystem.
Graph-Based Scam Detection: Leverage advanced graph algorithms (such as community detection and Graph Neural Networks/GNN embeddings) to map complex fraud rings and uncover hidden relationships between entities.
AI-Assisted Operational Workflows: Build intelligent, automated investigation pipelines using modern orchestration frameworks like LangGraph to streamline and accelerate scam detection operations.
Machine Learning & Predictive Modeling: Deploy robust classical machine learning approaches (including gradient boosting and bagging algorithms) to classify high-risk behavior with high precision and low latency.
Learning Outcomes
- Build predictive graph-based detection algorithms to unify data across domains
- Work closely with internal and external stakeholders to craft objectives and outcomes of detection systems
- Work with live end-to-end big data pipelines
Prerequisites
Must have:
- Experience with git and python
- Model building experience and principles (train/test split, metrics)
- Relational database concepts
Good to have:
- Experience with LLMs, langgraph, AWS, graph databases