• Architecting and Implementing Agentic Workflows • Design Autonomous Loops: Build stateful control loops (Perception -> Reasoning -> Action) where agents can interpret clinical data, plan necessary checks, and generate risk assessments. • Conversational Agentic workflow: Chatbot style user experience powered by LLM • GCP Integration: Leverage Vertex AI for model serving and LLM models for reasoning, optimizing for long-context windows to process complex medical history. • RAG & Grounding: Implement "Enterprise Grounding" using Vertex AI Vector Search to ensure all agent outputs are strictly cited against official medical literature (e.g., medical research and guidelines). • Data & Tool Integration • Structured Data Access: Build robust Python tools (Function Calling) that allow agents to securely query BigQuery or other datastores for patient demographics and symptoms. • Security & Compliance: Ensure all AI operations comply with HIPAA/GDPR standards, implementing strict PII masking and data governance within the Google Cloud environment. • System Reliability: Deploy agents on Vertex AI, Cloud Run, or GKE, ensuring low-latency inference and high availability for clinical users. • Agent monitoring: implement effective product environment AI Agent execution monitoring and alert system to ensure SLA, safety and security compliance • Leadership & Strategy • Mentorship: Guide a team of engineers in MLOps best practices (CI/CD for models, evaluation pipelines). • Evaluation: Implement "LLM-as-a-Judge" frameworks to automatically test agent accuracy against "Golden Datasets" of clinical cases.

ZeroFox protects organizations from external threats across the public attack surface, and we're building AI into how the company itself operates, not just what it ships to customers. You'll be the person who makes that real. This is a broad, hands-on engineering role. You'll assess business unit workflows, identify where automation creates real leverage, and build the systems that deliver it, whether that's an agent, a custom tool, an integration, or something simpler. You'll also evaluate whether enterprise data is ready for AI consumption and build the access patterns that close the gap. You'll make the judgment on the right approach for each problem, build it, and see it through to production. Some internal tooling may inform customer-facing capabilities over time, so you'll navigate that boundary as it comes up. What you'll do - Assess business unit workflows and identify where automation, agents, or custom tooling create real leverage - Design and build agents for internal operations: orchestration, tool-calling patterns, escalation paths, and human-in-the-loop gates - Build custom internal tools, integrations, and infrastructure that support business operations across functions - Evaluate enterprise data systems for AI readiness: quality, accessibility, governance, and latency. Build the access patterns that make enterprise data usable by AI systems. - Define the AI application lifecycle for internally-built systems: how artifacts get verified, tested, deployed, and maintained. Establish quality gates between prototype and production. - Define and enforce AI governance practices: agent permissions, escalation paths, decision audit trails, behavioral monitoring, and the boundary between autonomous action and human oversight - Build and operate internal AI infrastructure: self-service capabilities, deployment guardrails, observability, and cost management - Build tools and interfaces that make AI capabilities accessible to non-technical teams, translating business problems into working systems they can use directly

Join us as a Software Engineer, AI - This is an opportunity for a driven Software Engineer, AI to take on an exciting new career challenge - Day-to-day, you'll be engineering and maintaining innovative, customer centric, high performance, secure and robust solutions - It’s a chance to hone your existing technical skills and advance your career while building a wide network of stakeholders What you'll do In your new role, you’ll be working within a feature team to engineer software, scripts and tools, as well as liaising with other engineers, architects and business analysts across the platform. You’ll also be: - Producing complex and critical software rapidly and of high quality which adds value to the business - Working in permanent teams who are responsible for the full life cycle, from initial development, through enhancement and maintenance to replacement or decommissioning - Collaborating to optimise our software engineering capability - Designing, producing, testing and implementing our working software solutions - Working across the life cycle, from requirements analysis and design, through coding to testing, deployment and operations The skills you'll need To take on this role, you’ll need a background in software engineering, software design, and architecture, and an understanding of how your area of expertise supports our customers. You’ll also need: - Experience of working with multiple programming languages including Python and Web Development, such as NodeJS - Expertise in cloud platforms such as AWS, along with Dev Ops and CI/CD - Experience of problem solving, building to requirements and testing for both production and fast innovative proof of concepts - Expertise in Data Science, Generative AI, ML and LLMs, having RAG and Agentic experience would be beneficial - Proficient in implementing programming best practice, especially around scalability, automation, virtualisation, optimisation, availability and performance Hours 35 Job Posting Closing Date: 04/04/2026 Ways of Working:Remote First

• Develop and maintain APIs using the FastAPI framework. • Implement background tasks and asynchronous processing to optimize system performance. • Designing and building LLM-powered features and pipelines (RAG, agents, tool use, multi-step reasoning chains) • Context engineering — structuring prompts, managing context windows, optimizing token usage, implementing retrieval strategies to feed LLMs the right information at the right time • Working with LLM APIs (Claude, OpenAI, etc.) and orchestration frameworks • Building evaluation and testing frameworks for LLM outputs (accuracy, hallucination detection, regression testing) • Implementing guardrails, content filtering, and responsible AI patterns • Optimizing latency and cost for LLM-heavy workloads (caching, streaming, batching) • Write clean, efficient, and maintainable code that adheres to industry best practices.