Software Engineer – Agent Guardrails

Job not on LinkedIn

🔥 15 minutes ago

🌐 Germany, India – Remote

🇩🇪 Germany – Remote 🇮🇳 India – Remote

⏰ Full Time

🟡 Mid-level

🟠 Senior

🧑‍💻 Full-stack Engineer

Apply Now

Find Similar Remote Jobs

Set Up Email Alert For New Jobs

📊 Check your resume score for this job

Improve your chances of getting an interview by checking your resume score before you apply.

Upload Resume

SMB Advisory

WebsiteLinkedInAll Job Openings

11 - 50 employees

🤝 B2B

💼 Consulting

B2B • Consulting

SMB Advisory is a specialist firm of Chartered Accountants dedicated to providing specialist services to Financiers, Advisors, Directors, and Individuals. The company focuses on delivering strategic, practical, and effective accounting solutions tailored to the unique needs of their clients.

📋 Description

• Design and build low-latency runtime instrumentation across Node.js and Python to track AI agent execution paths in real time • Pioneer hooks and interceptors for emerging protocols such as MCP, LLM SDKs, web frameworks, and new execution transports • Engineer automated systems and harness tooling to detect, test, and automatically surface upstream SDK shifts before they impact runtime behavior • Enforce strict execution overhead budgets so guardrails do not compromise host agent execution speed • Own the language-runtime sensors and instrumentation layer that hooks into key packages, generates events for the policy engine, and enables real-time block decisions

🎯 Requirements

• Production mastery in Node.js and Python • Experience with dynamic module loading, monkey-patching, and low-level runtime interception in both Node.js and Python • Comfort bridging managed runtimes into compiled binaries such as Rust, C, or C++ • Fundamental understanding of memory management, thread safety, and dynamic linking • Deep understanding of non-blocking control flow, including asyncio and the Node event loop • Proven track record building resilient software on fast-evolving third-party SDKs and uncooperative runtimes • Diagnostic mindset for identifying race conditions, memory leaks, and silent failures across runtime boundaries