Build evaluation tooling: golden question sets, retrieval and answer-quality metrics, and regression checks in CI.
Expose retrieval through REST APIs and tool interfaces such as MCP so that AI agents can use it as a tool.
Run the services in production: tests, observability, latency, and cost-aware use of GPUs and LLM endpoints, including on-premises and air-gapped deployments.
Support onboarding of product teams and share findings through short write-ups and demos.
Qualifications
Required
5 to 8 years of professional Python backend engineering; REST APIs, SQL, async programming.
Has owned a production service or pipeline end to end, and can guide less experienced engineers.
Working knowledge of RAG and some hands-on experience: you have built or contributed to a retrieval or LLM-based application (a project, prototype or production feature) and understand chunking, embeddings, vector search and how to judge result quality. A strong backend engineer who has moved into RAG is a good fit; you do not need years of RAG-only experience.
Experience building document-processing or data-ingestion pipelines.
Containers and CI/CD; comfortable deploying to Kubernetes.
Disciplined testing and code review habits.
Collaborates with colleagues across India, the US and the UK, and communicates clearly in English.
Preferred
Production RAG experience: chunking and embedding choices, hybrid search, reranking, and how you measured quality.
LLM and retrieval evaluation (RAGAS-style or custom); embedding model selection or fine-tuning.
LLM endpoints (Azure OpenAI, Bedrock, self-hosted models); tool-calling, MCP or agent-framework integrations (a plus, not required).
Experience with technical documentation or semi-structured content.
Curiosity about test and measurement or telecom domains.
Experience using AI coding assistants (such as GitHub Copilot or Claude) in day-to-day development, with the judgement to review and test what they produce.
Careers Privacy Statement***Keysight is an Equal Opportunity Employer.***
Build and maintain the ingestion pipeline in Python: document parsing, chunking, embedding, indexing, versioning of sources and access control.
Improve retrieval quality: hybrid (keyword plus vector) search, reranking, query rewriting and metadata filtering, including for technical content such as manuals and reference documents.
Build evaluation tooling: golden question sets, retrieval and answer-quality metrics, and regression checks in CI.
Expose retrieval through REST APIs and tool interfaces such as MCP so that AI agents can use it as a tool.
Run the services in production: tests, observability, latency, and cost-aware use of GPUs and LLM endpoints, including on-premises and air-gapped deployments.
Support onboarding of product teams and share findings through short write-ups and demos.
Required
5 to 8 years of professional Python backend engineering; REST APIs, SQL, async programming.
Has owned a production service or pipeline end to end, and can guide less experienced engineers.
Working knowledge of RAG and some hands-on experience: you have built or contributed to a retrieval or LLM-based application (a project, prototype or production feature) and understand chunking, embeddings, vector search and how to judge result quality. A strong backend engineer who has moved into RAG is a good fit; you do not need years of RAG-only experience.
Experience building document-processing or data-ingestion pipelines.
Containers and CI/CD; comfortable deploying to Kubernetes.
Disciplined testing and code review habits.
Collaborates with colleagues across India, the US and the UK, and communicates clearly in English.
Preferred
Production RAG experience: chunking and embedding choices, hybrid search, reranking, and how you measured quality.
LLM and retrieval evaluation (RAGAS-style or custom); embedding model selection or fine-tuning.
LLM endpoints (Azure OpenAI, Bedrock, self-hosted models); tool-calling, MCP or agent-framework integrations (a plus, not required).
Experience with technical documentation or semi-structured content.
Curiosity about test and measurement or telecom domains.
Experience using AI coding assistants (such as GitHub Copilot or Claude) in day-to-day development, with the judgement to review and test what they produce.
Careers Privacy Statement***Keysight is an Equal Opportunity Employer.***