Senior Research Specialist, AI & ML

To make electronics, you need chips, wafers, transistors, reticles, and... To make these, you must see, test and manufacture them at scale—faster and better than ever before. That's where KLA comes in. Whether you're early in your career or an experienced professional, you'll solve complex challenges, work alongside brilliant minds and help shape the future of technology.

Group/Division
The Semiconductor Products and Customers (Semi PC) business unit designs, builds and sells KLA’s product portfolio to help chip manufacturers meet high-quality standards for technologies like AI, data centers, automotive and electronic devices. Our inspection, metrology, and analytics systems detect and monitor defects across logic and memory chips, specialty semiconductors, wafers, reticles, advanced packaging, process equipment and materials. We also develop technologies for specialty processes, IC substrates, and PCB manufacturing, including etch and deposition, component inspection, and imaging and analytics. At KLA, physics and data are at the core of everything we do. Our AI and Modeling Center of Excellence drives innovation across KLA products by combining expertise in physics modeling, high-performance computing (HPC) and AI. The group is comprised of top-tier physicists, machine learning engineers, HPC architects and data scientists that collaborate to develop scalable, cutting-edge solutions for our customers. We’re hiring across multiple disciplines—this is your opportunity to make a global impact.

What You'll Do

In this role, you will play a key part in advancing business priorities by delivering high-impact work across your area of expertise.

The work spans both applied and fundamental research. You will identify high-value research problems, formulate them mathematically, develop novel learning methods, build proof-of-concept systems, and conduct rigorous evaluations. Successful research outcomes will be translated into scalable solutions through collaboration with multidisciplinary engineering and domain teams.

This role is not focused on building conversational assistants or connecting foundation models through predefined workflows. The central challenge is to develop intelligent systems that interact with analytical tools, reason over heterogeneous engineering data, make sequential decisions, and improve through feedback. We are looking for researchers who view agentic AI as a learning and decision-making problem involving observations, actions, tools, feedback, uncertainty, and measurable outcomes. We value scientific rigor and sound judgment, selecting methods based on the structure of the problem rather than on current trends.

What You Will Work On

You will develop intelligent systems capable of:

- Reasoning over images, text, measurements, metadata, and domain knowledge.
- Planning and executing multi-step analytical workflows.
- Selecting and invoking specialized algorithms, models, and software tools.
- Examining intermediate results and deciding what action to take next.
- Learning from expert demonstrations, human feedback, and execution outcomes.
- Improving decision policies through reinforcement learning.
- Explaining decisions using grounded and verifiable evidence.
- Operating reliably in complex scientific and engineering environments.

Potential Research Areas

- Reinforcement learning for foundation-model reasoning and post-training.
- Multimodal agents for engineering analysis and decision support.
- Sequential decision-making for inspection and metrology workflows.
- Adaptive tools, algorithms, measurement, and data acquisition strategies.
- Learning from domain-expert feedback and verifiable execution outcomes.
- Reward design, agent trajectory learning and optimization.
- Human-agent collaboration and knowledge-grounded reasoning.
- Agent reliability, interpretability, and failure of recovery.
- Fine-tuning and evaluating LLMs and VLMs for specialized domains.

Day-to-Day Responsibilities

- Identify high-value opportunities for reinforcement learning, agentic AI, and multimodal reasoning in semiconductor inspection and metrology.
- Formulate scientific and engineering workflows as learning and sequential decision-making problems.
- Design intelligent agents that reason over multimodal data, use specialized tools and adapt their actions based on feedback.
- Develop and evaluate learning methods that improve agent reasoning, planning, tool use, and decision quality.
- Adapt foundation models to specialized scientific and industrial applications.
- Design rigorous experiments to assess performance, reliability, generalization, grounding, and failure recovery.
- Build reproducible prototypes and collaborate with engineering, infrastructure and domain teams to translate successful research into scalable solutions.
- Communicate research findings and collaborate with global multidisciplinary teams.
- Manage a team of engineers working in the space of AI

Minimum Qualifications

- PhD in Computer Science, Electrical Engineering, Artificial Intelligence, Machine Learning or a related discipline, with 3 to 5 years of relevant research or industrial experience; or a master’s degree with 5 to 10 years of relevant experience. Candidates with less experience may be considered if they demonstrate an exceptional research or engineering record.
- Strong foundation in at least one of the following: reinforcement learning; sequential decision-making; learning-based agents; foundation-model post-training; or planning and decision-making under uncertainty.
- Hands-on experience designing, implementing or evaluating intelligent agents involving planning, tool use, retrieval, memory, self-correction, multi-step execution, or learning from feedback.
- Working experience with LLMs, VLMs, multimodal foundation models, or generative foundation models.
- Sound understanding of reinforcement-learning concepts, including policies, rewards, value functions, exploration, credit assignment, policy optimization and evaluation.
- Ability to formulate real-world problems in terms of observations, states, actions, rewards, and sequential objectives.
- Strong foundation in deep learning, transformer architectures and representation learning.
- Hands-on experience developing, training, fine-tuning and evaluating models using PyTorch, JAX, TensorFlow or a comparable framework.
- Strong Python programming skills and the ability to develop maintainable and reproducible research software.
- Ability to formulate research and engineering problems mathematically.
- Experience designing controlled experiments, defining appropriate baselines, and selecting meaningful evaluation of metrics.
- Ability to analyze experimental results, identify failure modes, and draw evidence-based conclusions.
- Ability to follow rapidly evolving research and translate relevant ideas into working prototypes.
- Strong written and verbal communication skills, with the ability to work effectively in multidisciplinary and geographically distributed teams.

Preferred Qualifications

- Research or industry experience in reinforcement learning, agentic AI, multimodal learning, or foundation-model adaptation.
- Experience developing intelligent systems that learn from feedback, interact with tools, and operate over extended decision-making horizons.
- Experience training, fine-tuning, or evaluating language, vision-language, or multimodal foundation models.
- Experience developing robust evaluation frameworks for agent performance, reliability, and generalization.
- Experience applying machine learning to complex scientific, engineering, or industrial problems.
- Evidence of strong technical depth across modern machine learning, deep learning, and AI systems research.
- Evidence of team leadership in small or large organizations.

Things That Would Make Us Go “Wow”

- A demonstrated record of advancing the state of the art in reinforcement learning, agentic AI, multimodal learning, or foundation models.
- Experience building intelligent agents that learn from feedback, use tools effectively, and operate reliably in complex environments.
- Evidence of applying advanced AI methods to challenging scientific, engineering, or industrial problems with measurable impact.
- Experience defining rigorous evaluation methodologies for agent performance, reliability, and robustness.
- A track record of translating research innovations into scalable systems, products, or business outcomes..

About KLA

We provide advanced inspection tools, metrology systems, process solutions, and computational analytics that make electronics possible, tackling complex challenges. From electron and photon optics to machine learning and data analytics, we seek perfection at the most fundamental level of matter in the universe. If you want to make electronics that push industries forward and make the world a better place, join us.

Use of AI Statement

At KLA, our interviews seek to understand your individual skills, problem-solving approach and authentic thinking. To ensure a fair and consistent evaluation, the use of AI, recording tools or other technologies to generate, suggest or provide responses during interviews—whether virtual or in person—is not permitted unless explicitly approved in advance as part of a reasonable accommodation or invited by the interviewer. Use of these tools may interfere with our ability to evaluate your individual qualifications and affect your candidacy. KLA is committed to advancing innovation through responsible AI, and we value candidates who share this mindset.

We offer a competitive, family friendly total rewards package. We design our programs to reflect our commitment to an inclusive environment, while ensuring we provide benefits that meet the diverse needs of our employees.

KLA is proud to be an equal opportunity employer.