HelpwareThe DevSecOps Engineer is a foundational role in the our partner’s project, responsible for architecting and maintaining the high-security infrastructure required for an enterprise-grade Agent Development Platform (ADP). You will be tasked with building the secure "sandbox" where autonomous agents live, ensuring zero-trust connectivity and managing the specialized data stores that power agentic reasoning.
● Secure Data Infrastructure: Provision and optimize the platform's multi-layered data stack, including PostgreSQL (OLTP), ChromaDB/PGVector (VectorStore), and GraphDB (Neo4j) for relational knowledge mapping.
● Event Orchestration: Build and manage high-performance Event Streams (AWS SQS, Event Bridge, or Kafka) to handle agent-to-agent communication and Human-in-the-Loop (HITL) escalation triggers.
● Infrastructure as Code (IaC): Automate the provisioning of AWS cloud infrastructure and Kubernetes clusters using industry-standard tools.
● Security & Identity Integration: Implement WorkOS/IDP for B2B identity management and multi-tenancy, ensuring strict logical separation of tenant data and configurations.
● Network Hardening: Design and enforce strict egress firewalls and proxy-only outbound routing to prevent data exfiltration by autonomous agents.
Required Skills & Experience
● Cloud & Containers: Deep expertise in AWS, Kubernetes, and container security; proven experience with Firecracker MicroVMs is highly desirable.
● Database Management: Hands-on experience managing and scaling Vector Databases (Chroma, PGVector) and Graph Databases for AI applications.
● Security First Mindset: Strong understanding of zero-trust architecture, OWASP principles, and secure design for AI systems.
● Messaging & Events: Proficiency in managing event-driven architectures and streaming platforms for real-time data processing.
● Automation: Advanced scripting skills (Python, Bash) and a track record of implementing automated security gates in CI/CD pipelines.
● Workflow Orchestration: Experience with Temporal for managing long-running, multi-stage agentic workflows.
● Agentic Frameworks: Familiarity with LangSmith or Langfuse for observability and tracing of agent workloads.
¿Te interesa este puesto?