Agentic AI Engineer

Rimini Street; Hyderabad; On-site; INR 1,500,000 - 2,500,000; Full time

Job summary

Rimini Street in Hyderabad is seeking an Agentic AI Engineer to design cognitive architectures for AI agents within enterprise ERP systems. This role involves architecting agent behavior systems, developing decision frameworks, and enhancing multi-agent coordination.

The ideal candidate should have over 7 years of software engineering experience with proficiency in Python or Java, as well as demonstrable experience with AI and LLM-based systems. Collaboration in a diverse work environment is emphasized.

Qualifications

• 7+ years of software engineering with strong Python and/or Java proficiency.
• 3+ years working with AI/ML systems in production, including LLM applications.
• 1+ years designing autonomous AI systems.

Responsibilities

• Design cognitive architecture for AI agents interacting with ERP systems.
• Create decision trees for complex ERP tasks.
• Develop agent personas with domain-specific reasoning patterns.

Skills

Python proficiency

Java proficiency

AI/ML systems experience

Agent frameworks understanding

Tool-calling patterns in LLMs

Tools

Pydantic AI

LangChain

CrewAI

AutoGen

Job description

Role Overview

As an Agentic AI Engineer, you will design and build the cognitive architecture of AI agents that interact with enterprise ERP systems. Your work will focus on agent behavior systems: reasoning through complex multi‑step problems, selecting tools, coordinating with other agents, governing autonomy, and escalating to humans at the right moments. This role sits at the intersection of AI engineering and enterprise systems architecture, moving beyond prompt engineering to full agentic system design.

Key Responsibilities

Agent Cognitive Architecture

• Design the reasoning framework for decomposing complex ERP tasks into structured decision trees with governance checkpoints.
• Create agent personas and behavioral profiles for Rimini Solution domains (Finance, Procurement, Supplier Management, Expense, Support) with domain‑specific reasoning patterns, risk tolerances, and escalation triggers.
• Architect dynamic tool‑selection strategies where agents choose which MPC tools to invoke based on context, confidence, and task requirements.
• Design conversation state management for long‑running agent sessions spanning multiple interactions, tool calls, and human‑in‑the‑loop approvals.
• Develop confidence‑scoring frameworks that translate LLM output uncertainty into actionable governance decisions.

Multi‑Agent Coordination

• Design Agent‑to‑Agent communication patterns for collaborative workflows.
• Architect supervisor/worker agent hierarchies with delegation, progress monitoring, and result aggregation.
• Build conflict‑resolution strategies when agents produce contradictory assessments.
• Design shared context and memory patterns that allow agent teams to build collective understanding without redundant processing.

Autonomous Decision Governance

• Translate OPA policy definitions into agent‑actionable decision boundaries.
• Design autonomy progression models for incremental trust and error‑driven degradation.
• Build an explainability layer that articulates reasoning chains to human reviewers.
• Implement circuit‑breaker patterns at the cognitive level to detect competence boundaries and proactively escape.

Knowledge Integration & Institutional Memory

• Design retrieval‑augmented reasoning to pull relevant knowledge from vector stores, historical data, and approval pattern databases.
• Build feedback mechanisms that capture user corrections and turn them into future reasoning improvements.

Agent Evaluation & Quality

• Design evaluation frameworks for agent behavior, including reasoning quality and governance compliance.
• Build scenario‑based testing methodologies covering edge, ambiguous, and adversarial ERP scenarios.
• Define performance metrics: task completion rate, escalation accuracy, error rates, time‑to‑resolution, user override frequency.
• Own continuous improvement loops analyzing production behavior and refining architecture.

Required Experience

• 7+ years of software engineering with strong Python and/or Java proficiency.
• 3+ years working with AI/ML systems in production, including LLM‑based applications.
• 1+ years designing or building autonomous AI systems that reason, plan, use tools, and take actions.
• Experience with agent frameworks such as LangChain/LangGraph, Pydantic AI, CrewAI, AutoGen, Semantic Kernel, or equivalent.
• Hands‑on with tool‑calling/function‑calling patterns in LLMs (MCP, OpenAI function calling, Anthropic tool use).
• Understanding of evaluation methodologies for non‑deterministic systems.
• Experience building systems that handle human‑in‑the‑loop workflows.

Required Technical Skills

• Advanced proficiency in Python and Pydantic AI; fluency in LangChain/LangGraph, CrewAI, or AutoGen.
• Deep experience with Claude (Anthropic), GPT (OpenAI), or equivalent LLMs, including prompts, tool calling, streaming, structured output.
• Hands‑on with MCP (Model Context Protocol) and multi‑agent communication patterns.
• Knowledge of A2A protocols, vector search, reranking, context assembly, and citation.
• Experience with RAG architecture, TruLens, Ragas, or custom evaluation frameworks.
• Familiarity with durable execution (Restate/Temporal) and workflow orchestration integration.

Preferred Skills

• Experience with ERP systems such as SAP, Oracle EBS, JD Edwards, and familiarity with AP, GL, procurement processes.
• Java / Quarkus experience to work across Python agents and Java platform.
• Knowledge of reinforcement learning from human feedback, knowledge‑graph design, decision theory, or risk assessment frameworks.
• Experience in regulated industries requiring explainability and auditability.
• Strong technical writing to document architectural decisions, reasoning patterns, and governance frameworks.

Equal Employment Opportunity

Rimini Street is committed to creating a diverse and inclusive environment and is proud to be an Equal Employment Opportunity Employer. All qualified applicants will receive consideration for employment without regard to age, race, color, religion, national origin, sexual orientation, gender or gender identity, disability, protected veteran status, or any other characteristic protected by law.