Posted on: 01/07/2026
Summary :
Project AIRA is a hub-and-spoke AIOps platform that automates access security operations using agentic AI. Engineers and internal users interact in plain English; AIRA reasons, plans, and acts across integrated systems with full human-in-the-loop safety controls. The platform serves three personas : Assistant (internal users), Copilot (engineers), and Auditor (compliance).
Responsibilities :
- Build the Hybrid Agent core (Planner, Memory, Tool Registry, Executor) using LangGraph for stateful agentic workflows.
- Develop microservice connectors (Panorama, ServiceNow, Forward Networks, Splunk, Zscaler, IAM) with standard health check / capabilities / execute contracts.
- Build the MCP Adapter SDK : PII redaction (Presidio), risk tagging, rollback references, prompt injection defense, and token compaction.
- Implement UITL approval workflows, immutable audit logging (corrId, rollbackRef), and CAB-compliant write gating.
- Build RAG pipeline over internal SOPs/KBs using OpenSearch; PostgreSQL for system-of-record; Redis caching with TTL tiers.
- Deploy and operate on CosmosAI / Kubernetes; Prometheus metrics, distributed tracing, and CI/CD with container scanning.
Requirements :
- 5 years backend/platform engineering in Python-based distributed systems.
- Hands-on LLM application development : agents, RAG, function calling, output validation.
- Microservices design with typed contracts, resilience patterns, and API security (OAuth, mTLS, secrets management).
- Containerized deployment on Kubernetes or equivalent; CI/CD and DevSecOps practices.
Required Skills :
- AI / LLM : LangGraph, LangChain, GPT-4 / Claude API, RAG, Guardrails, structured function calling.
- Microservices : Python (FastAPI/Flask), MCP, OAuth / mTLS, Server-Sent Events (SSE).
- Security : Vault, PII redaction (Presidio), prompt injection defense, UITL, audit logging.
- Infrastructure : Kubernetes / CosmosAI, PostgreSQL, OpenSearch, Redis, Prometheus.
- Resilience : Retry / backoff, circuit breakers, idempotency keys, chaos testing, distributed tracing.
- CI/CD : Container scanning, policy-as-code, GitOps, automated integration tests.
Preferred Skills :
- Integrations : Palo Alto Panorama, Zscaler, ServiceNow, Splunk, Forward Networks, IAM.
Did you find something suspicious?
Posted by
Abdul Razzak
Last Active: NA as recruiter has posted this job through third party tool.
Posted in
AI/ML
Functional Area
Backend Development
Job Code
1650346