Agentic AI Architect

Advanced AI Operations (AIOps) & Agentic Knowledge Ecosystem

Project overview
The objective of this project is to develop and deploy a sophisticated AI ecosystem leveraging Expert Digital Engineer's capabilities to transition from traditional RAG (Retrieval-Augmented Generation) to an Agentic AI Framework. This system will integrate deep engineering knowledge, automate complex technical workflows, and provide decision support for Master Plans, Project Studies, and Operational Troubleshooting.

Technical scope & core capabilities

Agentic AI Architecture
- Multi-agent orchestration: Implementation of a multi-agent system using LangGraph to manage complex, non-linear workflows where specialized agents collaborate to solve technical problems.
- Agentic RAG & Graph RAG: Transition from vector-only search to Graph RAG, utilizing Knowledge Graphs to maintain relationships between complex engineering entities.
- Implementation of MCP (Model Context Protocol) to standardize how AI agents interact with external data sources and tools.
- Prompt engineering: Systematic development and optimization of prompt templates to ensure high-precision outputs for technical engineering queries.
Tool integration & automation
- Engineering tool invocation: Enabling the AI agents to programmatically trigger and interface with specialized industry software, including:
  - Hysys (Process Simulation)
  - OLGA (Transient Flow Simulation)
  - Ansys (Engineering Simulation)
- External dependencies: Management and procurement of necessary APIs and third-party middleware required to bridge the AI ecosystem with legacy engineering software.
Knowledge management & governance
- Interactive indexing: Development of a feedback loop where subject matter experts (SMEs) can correct or refine indexed data, improving the Knowledge Graph over time.
- Document control:
  - Versioning controls: Implementation of tracking to ensure the AI retrieves the most current revision of a technical document.
  - Narrow search (filtering): Advanced filtration layers to restrict search scopes to specific projects, dates, or document types.
- Observability: Integration of Langfuse for tracing, debugging, and monitoring the performance and latency of AI chains and agent decisions.

Functional applications & business deliverables

Engineering & project support
- Master plan development: Utilizing AI to synthesize data for the study and development of corporate Master Plans and large-scale projects.
- Technology evaluations: Automating the benchmarking and evaluation of new technologies against existing corporate standards.
- Project reviews: AI-driven support for eReview and HAZOP (Hazard and Operability Study) participation, providing instant access to safety standards and historical data.
Operational excellence & maintenance
- Incident investigation: Supporting major incident investigations by correlating telemetry data with historical reports via Graph RAG.
- CRM & troubleshooting: Providing real-time technical support for operations through CRM queries and automated troubleshooting guides.
- Regulatory & corporate compliance:
  - Automated generation and issuance of quarterly crude quality reports.
  - Systematic updates to corporate engineering standards.
  - Support for corporate program inspections (environmental, safety, etc.).

Key performance indicators

Reduction in retrieval time: Decrease in time spent searching for specific technical data across versioned documents.
Tool automation rate: Percentage of simulation tasks (Hysys, OLGA, Ansys) successfully initiated via AI agents.
Accuracy rate: Improvement in response precision through the use of Interactive Indexing and Graph RAG.
Operational efficiency: Reduction in turnaround time for CRM queries and crude quality reporting.

Requirements

Must have a minimum of 8+ years of relevant experience.

Position details

Location: Khobar, Saudi Arabia
Industry: Information Technology (IT) / Software
Experience required: 6+ years

Show