below marketmost similar roles pay hereabove market
This role pays more than
55%
of similar roles. Most pay
$152,568–$235,100
— the blue band above.
At the midpoint, this role pays about
$201k
versus about
$194k
for comparable roles.
Based on 240 similar postings.
Employer
About ASML
ASML is a leading supplier of lithography equipment, used by the world’s top chipmakers to print microchips that are increasingly powerful, fast and energy efficient.
ASML currently has
72 open roles
on FindRole.
Listed pay typically runs
$122,400–$174,175
across 52 roles with salary data.
AI Research and Development Engineer joins the team at ASML to work at the intersection of advanced artificial intelligence and semiconductor inspection technologies. This research-focused role involves owning end-to-end LLM and VLM development workflows, including data curation, model selection, training, evaluation, and deployment. The engineer will design adaptation strategies like supervised fine-tuning, instruction tuning, and preference optimization while building high-quality training datasets and synthetic-data generation pipelines. Key responsibilities include developing Agentic AI solutions with planning and memory, and establishing benchmarking frameworks for robustness, latency, and throughput. Required skills include Python, PyTorch, and experience with transformer-based models, distributed training, and multi-GPU environments. The role solves complex engineering problems by translating research concepts into functional prototypes for semiconductor inspection and other advanced manufacturing applications.
Own end-to-end LLM/VLM development workflows including data curation, model selection, training, and deployment.
Design and execute model adaptation strategies such as supervised fine-tuning, instruction tuning, and preference optimization.
Build and maintain high-quality training datasets using cleaning, labeling, and synthetic-data generation techniques.
Establish rigorous benchmarking frameworks to evaluate model quality, robustness, hallucination, latency, and cost.
Develop Agentic AI solutions that integrate models with tools, planning, memory, and workflow orchestration.
Apply deep learning techniques to solve complex engineering problems in semiconductor inspection.
Translate engineering requirements into model-development objectives and validate emerging technologies with customers.
Transition validated AI research and prototypes into reusable platforms and product applications.
What we're looking for
Master’s or PhD degree in Computer Science, AI, Machine Learning, Electrical Engineering, or a related field; equivalent practical experience may be considered.
Demonstrated hands-on experience training or fine-tuning LLMs and/or VLMs across the end-to-end model development lifecycle.
Practical experience with modern deep-learning frameworks and model-development tooling, especially Python and PyTorch.
Strong understanding of transformer-based models, tokenization, optimization, training dynamics, evaluation methods, and common model failure modes.
Experience working with distributed training or multi-GPU environments and diagnosing training issues such as instability, overfitting, and performance regressions.
Strong programming skills to independently design, implement, test, and evaluate AI algorithms and prototypes.
Experience developing Agentic AI systems, including tool-using agents, planning, memory, and multi-agent systems (preferred).
Experience with LLM/VLM adaptation techniques such as LoRA, QLoRA, supervised fine-tuning, or preference optimization (preferred).
LLMs
Reinforcement Learning
Python
PyTorch
JAX
Deep Learning
Distributed Training
Transformers
Data Filtering
Curriculum Learning
Preference Optimization
Model Steering
Synthetic Data Generation
Evaluation Methodologies
LLMs
Reinforcement Learning
Python
PyTorch
JAX
Deep Learning
Distributed Training
Transformers
Data Filtering
Curriculum Learning
Model Steering
Preference Optimization
Synthetic Data Generation
Evaluation Methodologies
LLM
RAG
Agent Orchestration
Knowledge Graphs
Small Language Models
Databricks
Snowflake
Distributed Systems
ML Platforms
Prompt Caching
Guardrails
Model Context Protocol
Human-in-the-loop