> Markdown version of [/jobs/ext/4741212-senior-software-engineer-ai-agents-cloud-platform](https://www.wearedevelopers.com/jobs/ext/4741212-senior-software-engineer-ai-agents-cloud-platform). Every page supports `.md` or `Accept: text/markdown`. Links point to the HTML versions so they work for humans too. Agent guide: [/agents.md](https://www.wearedevelopers.com/agents.md). --- # Senior Software Engineer - AI Agents & Cloud Platform - **Company:** castaigroupinc - **Location:** Spain - **Contract:** Permanent contract - **Skills:** Clean Code Principles, Application Programming Interfaces (APIs), Artificial Intelligence, Amazon Web Services, Microsoft Azure, Software as a Service, Computer Networks, Continuous Integration, Software Debugging, DevOps, Distributed Systems, Identity and Access Management, OAuth, OpenID, Performance Tuning, System Programming, YAML, Datadog, Google Cloud, Cloud Platform System, Autoscaling, Retrieval-Augmented Generation, Large Language Models, Prompt Engineering, AI Coding Agents, Agentic-AI, Build Management, AI Platforms, Kubernetes, Low Latency, Virtual Agents - **Published:** November 16, 2026 - **Apply:** https://www.jobleads.com/es/job/senior-software-engineer-ai-agents-cloud-platform--spain--ea424458907d89eab15fa81684cfd1173 ## About the Role * Strong software engineering fundamentals: you design clean abstractions, write maintainable code, and can justify your architecture choices, not just make something work. You\'re a generalist who is as comfortable debugging a distributed system as designing an auth flow or tuning an agent\'s behavior. * Production experience with Go is strongly preferred; candidates without Go should demonstrate strong systems programming skills in a comparable language. * Experience building or operating large-scale SaaS platforms in production: multi-tenant systems, high volumes, and the failure modes that only show up at scale. * Solid understanding of Kubernetes. Our entire platform runs on it, and the assistants we build answer questions about it. You should be at home with autoscaling, scheduling, and networking concepts. * You\'ve personally driven a complex project end-to-end. * Hands-on experience with AI coding agents and agentic development workflows. This is how we build, not just what we build. * You\'ve used observability tooling in production. * CI/CD and DevOps practices experience. * Strong English skills, both verbal and written. * Startup mindset: adaptable, proactive, and comfortable with ambiguity. What will make you stand out * You\'ve built agentic systems that shipped : MCP servers or clients, agent gateways, tool-use orchestration, RAG pipelines, evaluation frameworks, and you can talk about what broke in production and how you found out. This is the single biggest differentiator for our team. * Deep hands-on experience with one or more of AWS, GCP, or Azure, with real understanding of IAM, compute, networking, and their APIs, not just console-level usage. * Experience with LLM application quality: evaluations, prompt engineering at scale, latency and cost optimization of agent loops. * Knowledge of OAuth 2.0 / OIDC and multi-tenant authorization design. * Deep knowledge of EKS, GKE, or AKS internals. * Kubernetes or cloud-native OSS contributions. ## Description What you\'ll be solving A platform engineer opens their AI client of choice, connects it to CAST AI, and types: "Why did our spend jump 40% in eu-west last week?" Thirty seconds later they have an answer, pulled from live cluster state and cost data, with links back to the evidence. No dashboards. No YAML archaeology. Just a question and a correct, permission-checked answer. Your job is to make that interaction boring. Reliably, securely, at production scale. Behind it sits the hard part: an AI assistant that reasons over thousands of production Kubernetes clusters, a platform that safely exposes company infrastructure to external AI clients, and quality pipelines that tell us when the agent is wrong before a customer does. Get it right, and operators stop clicking through consoles entirely. Get it wrong once, and an agent gives a confidently wrong answer to someone managing production. CAST AI already runs Kubernetes infrastructure autonomously. Our automation makes thousands of cost-and-reliability decisions a day. Cast Pilot is the team building the intelligence layer on top of that: AI assistants for Kubernetes and the cloud itself . Our assistant already answers thousands of questions a month from hundreds of organizations. Now we\'re opening the platform up: a public, permission-aware MCP server so customers can plug CAST AI straight into their own AI clients and agents, deeper connectivity across all three major clouds (AWS, GCP, Azure), and a continuous push to make our agents smarter, better grounded, and more capable. This is greenfield AI platform work on top of battle-tested infrastructure. Most features ship in 1 to 4 weeks. This is a location-specific opportunity. We are currently accepting applications from candidates residing in the following European countries: Bulgaria, Croatia, Estonia, Greece, Hungary, Latvia, Lithuania, Poland, Romania, Slovakia, Cyprus, Malta, Finland, Sweden, Ireland, Moldova, North Macedonia, Montenegro, Slovenia, and Ukraine. Requirements * Strong software engineering fundamentals: you design clean abstractions, write maintainable code, and can justify your architecture choices, not just make something work. You\'re a generalist who is as comfortable debugging a distributed system as designing an auth flow or tuning an agent\'s behavior. * Production experience with Go is strongly preferred; candidates without Go should demonstrate strong systems programming skills in a comparable language. * Experience building or operating large-scale SaaS platforms in production: multi-tenant systems, high volumes, and the failure modes that only show up at scale. * Solid understanding of Kubernetes. Our entire platform runs on it, and the assistants we build answer questions about it. You should be at home with autoscaling, scheduling, and networking concepts. * You\'ve personally driven a complex project end-to-end. * Hands-on experience with AI coding agents and agentic development workflows. This is how we build, not just what we build. * You\'ve used observability tooling in production. * CI/CD and DevOps practices experience. * Strong English skills, both verbal and written. * Startup mindset: adaptable, proactive, and comfortable with ambiguity. What will make you stand out * You\'ve built agentic systems that shipped : MCP servers or clients, agent gateways, tool-use orchestration, RAG pipelines, evaluation frameworks, and you can talk about what broke in production and how you found out. This is the single biggest differentiator for our team. * Deep hands-on experience with one or more of AWS, GCP, or Azure, with real understanding of IAM, compute, networking, and their APIs, not just console-level usage. * Experience with LLM application quality: evaluations, prompt engineering at scale, latency and cost optimization of agent loops. * Knowledge of OAuth 2.0 / OIDC and multi-tenant authorization design. * Deep knowledge of EKS, GKE, or AKS internals. * Kubernetes or cloud-native OSS contributions. Responsibilities * Design and build the systems behind our AI assistant: agent flows, context retrieval, and the data connections that let it reason about live Kubernetes infrastructure. * Build and evolve our public, permission-aware MCP server : curated tools, identity, audit, and strict safety gates, so customers can use CAST AI from their own AI clients. * Develop and maintain cloud connectivity across AWS, GCP, and Azure, and the data foundations our agents build on. * Write production Go services that interact with cloud provider APIs, the Kubernetes control plane, and LLM APIs in real time. * Own features end-to-end: from design through implementation, testing, and production rollout (most projects ship in 1 to 4 weeks). * Debug novel problems: an agent giving a wrong answer, a cross-cloud permission edge case, a data freshness issue that only appears at scale. * Build the quality harness: evals, telemetry, feedback loops, that let us ship agent improvements without regressions. Contribute