AI / Search Architect for Technical Plan Development (RAG on-premise)

Search Type

Senior / Principal profile - Architect / Technical consultant with possible continuity as Tech Lead in the execution phase.

Brief Context

We are looking for a senior expert in Artificial Intelligence applied to document search to DEVELOP the complete technical plan of an AI project in an on-premise environment (Red Hat OpenShift / OKD).

The role does not start from a closed solution: there is a discovery and initial guidelines, but the goal is for the candidate to build, structure, and prepare an executable technical plan, which can then lead in its implementation.

It is not a position of immediate execution or research, but of applied architecture and deep technical planning.

Main objective of the role

Design and document the comprehensive technical plan for an AI MVP based on hybrid RAG, focused on document search, prioritizing speed, correctness, and traceability, within on-premise constraints.

Project Scope

• Existing system:
  • Backend Laravel / REST API
  • DB: PostgreSQL
  • Current Search: Typesense
  • Infrastructure: Red Hat OpenShift / OKD (on-premise)
• AI Project:
  • Smart search engine (not a generic chatbot).
  • Hybrid RAG architecture (keyword + vectorial).
  • Extractive responses, with sources.

Key Responsibilities

Phase 1: Development of the technical plan

• Define the complete technical architecture of the AI MVP:
  • components,
  • flows
  • Responsibilities,
  • design decisions
• Design the document indexing pipeline:
  • what is indexed,
  • when,
  • how
• Define:
  • embedding strategy,
  • hybrid search,
  • scoring and ranking,
  • managing permissions and context.
• Establish criteria for:
  • performance,
  • correctness,
  • success metrics.
• Identify technical risks and mitigations.
• Deliver a clear, documented, and executable technical plan:
  • architecture,
  • phases,
  • dependencies,
  • explicit assumptions.

Potential Phase 2: Execution

• Lead the technical implementation of the defined plan.
• Support the development and infrastructure team.
• Adjust the design based on real execution feedback.

Mandatory Requirements

• Proven experience in:
  • RAG (Retrieval Augmented Generation).
  • Hybrid search (keyword + vectorial).
• Experience in Red Hat infrastructure:
  • OpenShift / OKD.
  • On-premise deployments.
• Demonstrable ability to:
  • design architectures,
  • document technical plans,
  • make trade-off decisions.

Desired Requirements

• Complex documental / administrative systems at large scale (hundreds of thousands of documents).
• Public sector or regulated environments.
• Search engines (Typesense, Elastic, OpenSearch).
• On-prem Vector DBs, OCR, PDFs.
• Definition of metrics (Recall@K, latency).

Profile sought

• Senior architect, very pragmatic.
• Strong focus on design and planning.
• Able to say what NOT to do with AI.