We use cookies. Find out more about it here. By continuing to browse this site you are agreeing to our use of cookies.
#alert
Back to search results
New

AI Intern

Advantest America
United States, Texas, Austin
Mar 26, 2026

Description
Advantest America, a global leader in Semiconductor Test and Measurement, is seeking a motivated and innovative engineering student to explore cutting-edge applications of machine learning and generative AI. This internship provides hands-on experience working with emerging
AI systems and integrating them into Advantest's advanced testing
platforms.

Location: Austin, TX or San Jose, CA (headquarters)

Role Overview
In this role, you will contribute to research and prototyping efforts focused on LLM-powered reasoning and evaluation systems. You will explore how retrieval-augmented generation (RAG) and agentic workflows can be used to analyze, compare, and assess complex technical content at scale. The internship emphasizes building AI systems that support decision-making, qualitative judgment, and structured feedback in real-world engineering and research environments.

You will work with unstructured and semi-structured documents, design multi-step reasoning pipelines, and evaluate system behavior against domain-specific expectations and constraints.

Key Responsibilities




  • Design and implement multi-step agentic workflows for analyzing and evaluating technical content.
  • Develop RAG-based pipelines that combine internal documentation and reference materials with LLM reasoning.
  • Build AI agents capable of:

    • Comparing proposed ideas or approaches against known solutions or baselines
    • Identifying conflicts, gaps, redundancies, or lack of novelty
    • Producing structured assessments and constructive feedback


  • Experiment with prompting strategies, planning, reflection, and tool usage to improve reasoning quality and consistency.
  • Evaluate and iterate on system performance using qualitative and semi-quantitative metrics.
  • Collaborate with engineers and researchers to translate ambiguous evaluation criteria into actionable AI workflows.

Applied = 0

(web-bd9584865-vpmzc)