Senior AI Memory Infrastructure Engineer

ByteDance; San Jose (CA); On-site; USD 212,800 - 450,000; Full time

Benefits offered by this job

Medical, dental and vision insurance

401(k) with company match

Paid parental leave

Disability coverage

17 paid personal time days (prorated)

Job summary

ByteDance is seeking a Senior Software Engineer for the AI Agent Memory Infrastructure in San Jose. You will design and evolve a unified memory platform supporting long-term, conversational, and task-oriented memory to power next-generation AI agents.

You will optimize pipelines for low latency and high availability, work on multimodal memory processing, and collaborate with model and product teams to productionize memory capabilities while balancing performance, cost, and safety.

Qualifications

• Bachelor’s degree or higher in Computer Science, AI, Data Science, or related fields.
• Proficient in Go, Python, or C++, with solid coding standards.
• Strong knowledge of LLM applications, embeddings, RAG, context engineering, and memory systems.

Responsibilities

• Design, build, and evolve memory infrastructure for AI agents and long-term memory.
• Architect and optimize memory pipelines for low-latency, high-availability environments.
• Explore challenges at the intersection of LLMs, context engineering, and data management.
• Design memory models and workflows for multimodal data (text, image, audio).
• Collaborate with model, application, and platform teams to productionize memory capabilities.

Skills

Distributed systems

Databases

Information retrieval

AI infrastructure

Go

Python

C++

LLMs

Context engineering

Education

Bachelor’s degree or higher in CS/AI/Data Science

Job description

ByteDance is seeking a Senior Software Engineer for the AI Agent Memory Infrastructure in San Jose. You will design and evolve a unified memory platform supporting long-term, conversational, and task-oriented memory to power next-generation AI agents.

You will optimize pipelines for low latency and high availability, work on multimodal memory processing, and collaborate with model and product teams to productionize memory capabilities while balancing performance, cost, and safety.