2025 Summer Intern, On-Device AI Agents Research Intern

Lab Summary: 

Samsung Knox™ (https://www.samsungknox.com/) is Samsung’s guarantee of security, and a secure device gives you the freedom to work and play how, where, and when you want. Samsung Knox consists of a highly secure platform built into a variety of Samsung devices, including Samsung’s mobile phones and laptop computers.

Come join the Samsung KNOX team and help us define and develop the future core technologies for Samsung devices and services!

Position Summary:

Samsung Research America Advanced Knox R&D Team is looking for an intern to contribute to the development of on-device AI agents optimized for enterprise productivity and security solutions. This role will focus on designing, optimizing, and deploying AI/LLM models for resource-constrained devices such as mobile phones.

Position Responsibilities:

• Assist in the design, development, and deployment of AI agents and frameworks tailored for on-device environments.
• Knowledge of context-aware systems, multi-agent systems, and decision-making frameworks for security and productivity driven enterprise applications.
• Develop proof-of-concept applications and assist in the performance evaluation of on-device AI models for B2B use cases.
• Collaborate with cross-functional teams to evaluate the performance, robustness, and security of AI agents in real-world enterprise scenarios.
• Work with mobile frameworks such as LiteRT, ExecuTorch, Mediapipe LLM, and similar technologies to deploy LLMs on mobile.
• Work on optimizing LLM-based applications for mobile devices, focusing on minimizing latency, power usage, and memory footprint.

Required Skills:

• Enrolled in a Ph.D. or Master's program in computer science or related areas. Hands-on experience in LLM model development, particularly for on-device or resource-constrained environments.
• Familiarity with NLP, and LLM fine-tuning, optimization techniques such as model pruning and model distillation.
• Experience with mobile AI frameworks such as LiteRT, ExecuTorch, Mediapipe LLM.
• Strong coding skills with prototyping experience using popular machine learning frameworks (e.g., PyTorch, TensorFlow).

Preferred Skills:

• Experience in deploying on-device LLM models.
• Experience with on-device deployment of AI agents capable of performing productivity and security tasks with minimal computational resources.
• Strong problem-solving skills with a focus on building efficient, real-world solutions for enterprise productivity and security applications.