Research Engineer – Multimodal AI & Interaction
Core
Research, design, and build next-generation multimodal AI capabilities for intelligent consumer devices, spanning speech/audio, vision, language, touch, gesture, and motion.
Role type
Senior IC research engineer (multimodal AI & interaction)
Builds
Prototype-driven multimodal AI systems for smartphones, wearables, and emerging devices
Domain
Consumer electronics + Multimodal AI (speech, vision, NLP, generative AI)
Deliverable
production ML models
Required skills
Deep learning, foundation models, multimodal learning, speech/audio processing, computer vision, NLP, generative AI, LLMs/VLMs, agentic AI, multi-agent systems, end-to-end ML development, model optimization for resource constraints
Preferred skills
Experience with LLMs/VLMs, agentic AI, multi-agent systems, context-aware intelligent systems, TypeScript, Kotlin, C++, Java
Technologies
PyTorch, TensorFlow, JAX, Python
Responsibilities
Design, train, fine-tune, and evaluate deep learning and foundation models for multimodal perception, reasoning, interaction, and generation; Identify high-impact research problems and own projects from early exploration through model development, evaluation, and product-oriented validation; Develop new approaches for reasoning across modalities to enable more natural, adaptive human-machine interaction; Rapidly prototype research ideas, reproduce state-of-the-art methods, and evolve promising concepts into robust demonstrations; Build datasets and evaluation methodologies, run rigorous experiments, and establish metrics reflecting both model performance and real interaction quality; Optimize models for real-world constraints (latency, memory, power, privacy, robustness, on-device vs. cloud)
Seniority
Senior, hands-on IC
