The posting
We are seeking an experienced AI/ML Engineer to lead the design, development, and scaling of advanced AI/ML solutions across our analytics platform in the manufacturing and semiconductor sectors. This high-impact role combines deep expertise in classical machine learning with cutting-edge Generative AI capabilities to deliver production-grade systems for anomaly detection, predictive maintenance, market intelligence, automated test plan generation, and expert-level customer support.
You will own end-to-end AI/ML initiatives — from numerical sensor/test data modeling to unstructured text processing and LLM-powered workflows — in a high-stakes, regulated industrial environment where precision, reliability, hallucination mitigation, and risk minimization are mandatory. This is a hands-on senior position requiring both architectural knowledge and strong implementation skills.
- Lead the architecture and continuous improvement of unified AI/ML capabilities, integrating classical ML models with Generative AI platforms (primarily AWS Bedrock) to support mission-critical applications in semiconductor manufacturing and risk analytics.
- Design and implement robust anomaly detection and predictive maintenance systems using classical ML algorithms (XGBoost, Scikit-learn) on real-time sensor and test data, while incorporating drift detection and model monitoring to maintain long-term reliability.
- Build and scale RAG pipelines and agentic workflows for high-precision tasks, including automated generation of manufacturing test plans from historical test data/measurement instrument records, with strong emphasis on accuracy, hallucination reduction, and risk controls.
- Develop intelligent summarization and information extraction pipelines that process thousands of scraped news articles, press releases, and open-source intelligence into concise, actionable market intelligence reports, leveraging techniques such as intelligent chunking, semantic filtering (embeddings + k-NN), map-reduce patterns, TF-IDF augmentation, and agentic orchestration.
- Own the development and maintenance of a customer-facing GenAI Q&A chatbot that provides deep, domain-specific insights into semiconductor manufacturing risks based on sensor measurements and test plans.
- Tackle diverse classical ML problems (regression, classification, clustering, time-series forecasting) and integrate them with GenAI components when hybrid approaches deliver better outcomes.
- Apply NLP techniques — including classical recurrent architectures (RNNs/LSTMs) and modern LLM-based methods — to extract insights from unstructured sources (market reports, operational logs, competitor pricing data).
- Collaborate with MLOps, data engineering, domain experts, and product teams in an Agile/Scrum environment to iterate models, conduct rigorous validation, ensure CI/CD, observability, versioning, and automated testing for all AI components.
- Perform advanced model evaluation, hyperparameter tuning, feature engineering, bias/risk assessment, and ethical AI practices, with particular attention to imbalanced datasets, concept/data drift monitoring, and production reliability.
- Contribute to large-scale data pipeline enhancements using tools like Apache Spark, vector databases, and distributed processing patterns.
- Stay current with advancements in classical ML, GenAI (RAG, agentic systems, multi-agent frameworks), responsible AI, and industrial analytics; proactively propose innovations that drive measurable business value.
Must-have qualifications
- Master's degree in Machine Learning, Computer Science, Data Science, Statistics, Quantitative Mathematics, or a closely related field.
- 4+ years of professional experience as a Machine Learning Engineer / AI Engineer (or equivalent), with a proven track record of independently owning end-to-end development, validation, and production deployment of both classical ML and GenAI/LLM-based systems.
- Strong hands-on expertise in classical ML frameworks (Scikit-learn, XGBoost) and deep learning/NLP (TensorFlow/PyTorch, RNNs/LSTMs)
- Practical experience building RAG architectures, prompt engineering, knowledge base curation, vector database optimization (embeddings tuning, hybrid search), and agentic workflows (LangChain/LangGraph, CrewAI, Bedrock Agents, or equivalent).
- Demonstrated success developing scalable summarization/information extraction pipelines for large document sets and production-grade anomaly detection/predictive models on numerical/time-series data.
- Proficiency in production-grade Python, clean code practices, Git, testing, CI/CD, and MLOps best practices (model monitoring, drift detection, automated retraining).
- Solid experience with AWS Bedrock (Knowledge Bases, custom models, Lambda/Step Functions for orchestration) or comparable GenAI platforms.
- Familiarity with Agile/Scrum, sprint-based delivery, cross-functional collaboration, and rigorous QA/validation of ML/GenAI systems (evaluation metrics, bias/risk assessment).
- Fluency in English, including technical terminology.
Strongly preferred
- Domain exposure to manufacturing, semiconductors, sensor-based analytics, test/measurement instrumentation, or industrial risk analytics.
- Hands-on experience with Apache Spark for large-scale processing and distributed computing.
- Prior work integrating classical ML with GenAI (e.g., hybrid pipelines, using classical models for filtering/reranking in RAG).
- A portfolio or demonstrable projects showing innovative, production-impactful solutions combining classical ML and Generative AI in real-world settings.
- Experience with the Model Context Protocol (MCP) for building standardized, secure integrations between LLMs/agentic systems and external data sources, tools, or enterprise services (e.g., connecting to databases, APIs, or knowledge repositories in a protocol-driven rather than custom-coded manner).
Careers Privacy Statement***Keysight is an Equal Opportunity Employer.***
