Memory and Personalization Intern

Responsibilities

• Design and prototype hybrid memory retrieval systems combining vector search and structured knowledge graphs
• Develop long-term memory models from user interaction signals
• Build evaluation pipelines for memory relevance, personalization accuracy, and latency
• Optimize memory systems for low-latency and on-device deployment
• Collaborate with system, platform, and product teams to integrate memory into assistant workflows

Requirements

• Currently pursuing a PhD in Computer Science or equivalent field
• Knowledge of Long-term memory, LLM, ML, and AI
• 2+ years of ML engineering experience with building end-to-end systems in production
• Strong hands-on experience with ML systems through research, internships, or production
• Experience with retrieval systems including vector databases, ranking, and search
• Familiarity with graph databases or knowledge graphs
• Experience with large language models and embedding-based retrieval
• Strong Python skills and familiarity with PyTorch or TensorFlow

Preferred Qualifications

• Experience with memory systems or knowledge representation
• Knowledge of GraphQL and modern API patterns
• Experience with recommendation systems or search
• Background in information retrieval or ranking systems
• Knowledge of model optimization and quantization
• Strong knowledge of design patterns and scalable architecture
• Experience with CI/CD platforms like Jenkins, GitHub Actions, or CircleCI
• Hands-on experience with Web applications/services development

About the Company

Advanced Intelligence Lab at Samsung Research America is working towards a vision of a world-class voice assistant that provides a unified AI experience across multiple devices including Galaxy Smartphones, Galaxy Watch, and smart home appliances.