Senior LLM Pipeline Engineer – AI Architect and Engineer for Generative AI Systems - FT - Europe, LATAM, NA, Asia

About the role:

We're seeking a Generative AI Engineer with deep experience orchestrating chain-like LLM workflows involving multiple interdependent tasks, agents, and external toolchains. Your systems should show robust sequential logic, checkpointing, error handling, and task memory management.

Requirements:

• 8+ years of experience in software engineering with advanced Python
• Direct experience engineering strong, deterministic prompts
• Designing agent sequences or pipelines for structured data flows
• Familiarity with LangChain Expression Language (LCEL), TaskGraph, or similar sequential LLM frameworks
• Strong debugging skills for tracing agent state, outputs, misrouting, and logical breakdowns across multistage workflows
• Experience integrating and managing workflows with AWS Bedrock (Claude, Titan, Bedrock Prompt Management, etc.)
• Skilled in Docker-based deployments, GitHub CI/CD pipelines, and .env-based environment isolation in GCP and AWS
• Proven ability to manage memory, agent state persistence, and error recovery in long-running LLM workflows
• Experience with long-context prompt management, summarization, and multi-stage chaining logic
• Demonstrated integration experience with external APIs and knowledge sources (e.g., MedPaLM, PubMed, UpToDate)
• Proficient with markdown-based sectioning, intermediate checkpoint saving, and stateful reruns
• Experience implementing staged pipelines with persistent folders and resumable execution flows
• Skilled in integrating patient intake forms and user-defined health concerns into prompt generation logic, with support for custom temperature/language settings
• Familiarity with LaTeX post-processing, citation formatting, and automated error mitigation for LLM-generated documents
• Experience building robust error-handling and retry systems for LLM pipelines (e.g., failed chapter generation, invalid citations, broken biomarker references)
• Demonstrated experience with prompt benchmarking, hallucination mitigation, and outline compliance across Claude, Gemini, o1/o3-mini, and other models
• Hands-on experience deploying and managing LLM pipelines on Google Cloud Platform (Cloud Run, Firestore, Cloud Functions, batch jobs)

Ideal to Also Have:

• Background in healthcare informatics or structured clinical data
• Experience deploying models via AWS Batch, with attention to permissions, logging, and cost optimization
• Strong documentation and sprint-tracking habits using Notion or ClickUp
• Experience contributing to robust, test-driven PR-based engineering workflows