AI Engineer - Remote (Spain)

Company Description At Keystone AI , we carve our own path. Trusted by Governments and world-leading pharma institutions and funded with $70m, we are on a mission to harness the power of connected data and bring precision medicine to life globally. The key to saving human lives is providing researchers across the globe with the right data to develop more precise drugs faster. Generating large amounts of biomedical data has become relatively straightforward. The challenge now is finding, accessing and analysing this sensitive data stored across thousands of disconnected locations. It is frustrating and painfully difficult for researchers. Keystone AI has created a patented, federated technology that brings analysis and computation to where data resides. This enables researchers to run analyses on multiple, distributed datasets in-situ, avoiding the risky movement of highly-sensitive data. Our Federated AI Platform is trusted by federal agencies, national health systems, and global pharma to securely unlock sensitive biomedical and real-world data at scale. We help federal health institutions modernize data utilization and analytics in highly protected environments, supporting mission-critical federal programs and transforming how secure health data drives public health and research. Our technology is unmatched and momentum is on our side - and we need an exceptionally strong team to help boost Keystone AI to incredible new heights. Your Impact As an AI Engineer, you will be at the forefront of bridging the gap between massive, decentralized biomedical datasets and life-saving insights. You won't just be building "standard" models; you will be co-architecting AI that functions within a federated ecosystem, where privacy and security are paramount. You will own the development and deployment of machine learning models—ranging from Large Language Models (LLMs) for clinical note extraction to predictive analytics for genomic research—ensuring they are optimized for our unique federated architecture. Your work will directly empower researchers to identify drug targets and disease biomarkers faster than ever before, transforming how the world approaches precision medicine. Your Role and Responsibilities Agentic Architecture & Orchestration Design and implement autonomous AI agents using frameworks like LangGraph, CrewAI, or AutoGen to handle complex, multi-step scientific queries. Develop sophisticated reasoning loops (e.g., ReAct, Plan-and-Execute) that allow agents to decompose high-level research goals into actionable sub-tasks. Build and optimize Advanced RAG (Retrieval-Augmented Generation) pipelines that integrate structured clinical data and unstructured scientific literature. Tool-Use & Execution Create and maintain "tools" for AI agents, enabling them to safely interface with Keystone AI ’s federated APIs, SQL databases, and bioinformatic execution engines. Implement secure, sandboxed code-interpreter capabilities, allowing agents to write and execute Python or R code for data visualization and statistical analysis. Fine-tune LLMs for specific function-calling and tool-use accuracy within the life sciences domain. System Reliability & Guardrails Develop robust evaluation frameworks (LLM-as-a-judge) to measure agentic performance, truthfulness, and safety in a clinical context. Implement "Human-in-the-loop" (HITL) patterns to ensure high-stakes scientific decisions are always reviewed by domain experts. Partner with Security teams to ensure agents operate within strict data privacy boundaries, preventing prompt injection or unauthorized data egress in federated nodes. Collaboration & Scaling Work with Product and UX teams to design intuitive interfaces for interacting with agentic systems (e.g., conversational research assistants). Scale agentic workloads in production using Kubernetes, ensuring low-latency reasoning and efficient token usage. Requirements Education: BSc/MSc in Computer Science, Artificial Intelligence, Machine Learning, or a highly quantitative field (PhD preferred). Experience: 2+ years of hands-on experience as an AI or ML Engineer, building a validated real product, ideally within a product-led biotech, health-tech, or SaaS company. Technical Stack: Deep proficiency in Python and Typescript and standard ML frameworks (e.g., Langfuse, PyTorch, TensorFlow, JAX, Scikit-learn). NLP/LLM Expertise: Proven experience working with Large Language Models, including fine-tuning, and RAG (Retrieval-Augmented Generation) architectures. Cloud & Infrastructure: Familiarity with AWS/Azure/GCP and experience deploying models in Docker/Kubernetes environments. Domain Knowledge: It is a plus to have experience working with biological, genomic, or clinical data is a significant advantage. Autonomy: A self-starter mindset with the ability to navigate ambiguity and drive AI projects from concept to production without constant oversight. What We Offer Keystone AI is synonymous with a dynamic work culture that encourages both pe