Agentic AI — PhD Intern (Modeling Agent)
$51 - $56 USD hourly
Originally posted 6 November 2026 by the employer.
About the role
This PhD internship focuses on building an agentic layer for adaptive engineering intelligence systems where measurement, simulation, and machine learning converge. You will design, build, and evaluate an auditable agentic flow that configures modeling and optimization workflows with minimal manual configuration.
What you'll do
- Build an agentic flow that configures data preparation, model training, and optimization for new engineering use cases.
- Develop tool integrations connecting the agent to simulation, data, optimization, and analysis systems.
- Implement autonomous data analysis to characterize unfamiliar datasets and select appropriate setups.
- Build explainability into the flow, covering configuration rationale, model dependencies, and confidence scores.
- Design guardrails including validation before expensive runs, bounded actions, and compute budgets.
- Assemble a benchmark of representative engineering use cases to measure task-completion reliability and result quality against an expert baseline.
What you'll need
- Currently pursuing or recently completed a PhD in Machine Learning, Computer Science, Electrical or Electronic Engineering, Applied Mathematics, or a related technical field.
- Strong Python and software engineering skills, including modular code, version control, and reproducible pipelines.
- Hands-on experience building LLM agents, including tool and function calling, orchestration, and context management beyond prompt-level experimentation.
- Demonstrated rigor in evaluating non-deterministic systems, including benchmark design, held-out test sets, automated scoring, and result interpretation.
- A portfolio of agent or LLM systems you have built, including code and evaluation methodology.
- Working knowledge of machine learning workflows, including data preparation, model selection, training, and optimization.
Nice to have
- Industry experience shipping agent or LLM-based systems to real users.
- Experience with multi-agent systems, agent frameworks, and MCP or comparable tool-integration protocols.
- Background in optimization, design of experiments, or surrogate modeling.
- Exposure to RF, microwave, or electromagnetic engineering problems.
- Experience with retrieval-augmented systems or structured generation.
- Familiarity with C++ for integration with internal engineering tooling.
Skills: LLM, agentic AI, Python, optimization, simulation
More semiconductor jobs: Keysight Technologies jobs · EDA tools engineer jobs
This role has been open 0 days — well below the 22-day median for EDA/CAD Engineering roles.
EDA/CAD Engineering · AI EDA Design
|
Open roles in category
99
|
Median days open
22 d
|
Category median salary (all employers)
$209k
|
See the full market breakdown ▾Category comparison, skills in demand, and who else is hiring
| Metric | Keysight Technologies | All employers we track in this specialty (99 roles · 28 employers) |
|---|---|---|
| Open roles in this specialty | 12 | 99 |
| Open roles in the wider Physical Implementation family | 47 | 1615 · 117 employers |
| Median days open | 25 d | 22 d (+3 d vs this employer) |
| Median salary (USD postings) | — | $209k |
Skills observed across this category: LLM, agentic AI, Python, optimization, simulation
Who's hiring in this category
- NVIDIA · 12 open roles · median 40 d
- NXP Semiconductors · 10 open roles · median 95 d
- Keysight Technologies (this employer) · 9 open roles · median 37 d
- STMicroelectronics · 8 open roles · median 3 d
- Synopsys · 7 open roles · median 15 d
- Cadence Design Systems · 6 open roles · median 71 d
How we counted: 99 open EDA/CAD Engineering (AI EDA Design) roles from 28 employers tracked in the SemiconductorJobs index, counted 6 Nov 2026. Specialty figures count only roles carrying this exact specialty label, so an employer's related work in neighbouring specialties is not included there — it is counted in the wider Physical Implementation family row. Figures refresh nightly.