AI Engineer | Self-Hosted AI Assistant, RAG & Agentic Automation

We're hiring an experienced AI engineer on behalf of an Australian business consultant who is building a private, self-hosted AI assistant for their consultancy. The goal is to create a ChatGPT-like experience that can securely access business knowledge, work with documents, connect to existing tools and autonomously complete routine tasks. This is not a basic chatbot or API wrapper project. We're looking for someone who understands how to bring together local LLMs, retrieval, AI agents and business automation into a production-ready system. *What We're Building The assistant will run locally on an NVIDIA DGX Spark (128GB unified memory) and be accessible through a web interface. The first version will focus on six key areas: 1. Private AI Chat & Local Models A polished, multi-user chat experience powered by locally hosted open-weight models, with the option to call external models such as Claude or GPT for more demanding tasks when explicitly authorised. 2. Business Knowledge & Document Retrieval (RAG) The assistant should securely access thousands of business documents, understand their contents and provide accurate, source-cited answers. This includes indexing shared folders, keeping information up to date and ensuring users only retrieve documents they're authorised to access. 3. AI Agents & Skills Agents that can use tools and perform real business tasks, such as: Preparing documents from existing templates and client records. Processing invoices and generating structured reports. Organising and filing documents. Extracting information from emails and attachments. Following defined business processes across different tools. We want reusable skills and workflows that can be added or modified without extensive development. 4. Proactive Automation The ability to schedule agents to work independently, monitor incoming information, prepare daily summaries and identify outstanding tasks. Agents should have clear permissions, maintain activity logs and request approval before taking sensitive actions. 5. Integrations Initial integrations will include Gmail and Dropbox, with a framework that makes adding other business applications straightforward. 6. Security, Privacy & Reliability Privacy is central to this project. Business data should remain on local infrastructure unless external access is explicitly authorised. The system will need proper user permissions, audit trails, controlled network access, encrypted backups and a reliable deployment and recovery process. *Who We're Looking For We're looking for a AI/backend engineer with proven experience deploying AI systems used by real people. You should have practical experience in several of the following areas: Local LLM deployment: Ollama, vLLM, NVIDIA GPUs, model optimisation and inference performance. RAG and document intelligence: Embeddings, vector databases, pgvector, hybrid search, document parsing and source citations. AI agent frameworks: Hermes Agent, OpenClaw, LangGraph or comparable frameworks involving tool use and persistent memory. Backend engineering: Python, PostgreSQL, APIs, background jobs, task scheduling and integrations. Infrastructure: Linux, Docker, networking, self-hosted applications and deployment automation. Security: Authentication, role-based access control, data isolation and secure handling of confidential information. Experience with MCP, OpenRouter, LiteLLM, OAuth, Gmail/Dropbox integrations or NVIDIA DGX hardware would be advantageous. More important than knowing every framework is knowing how to choose the right tools and make them work together. As part of your application please let me know how many followers labourcare has on instagram. We strongly prefer building on established open-source projects rather than developing custom infrastructure unnecessarily. *The Right Person The ideal candidate has already built something similar, whether that's a self-hosted AI assistant, an enterprise RAG platform, an autonomous agent system or an AI-powered business automation application. You'll be comfortable making architectural decisions, troubleshooting difficult problems and explaining technical trade-offs to a non-technical client. We're particularly interested in developers who use AI coding tools effectively while maintaining good engineering practices, including testing, documentation and version control. This is a hands-on engineering role. We need someone who can design the solution and build it, not just advise on architecture. There is potential for an ongoing monthly retainer to maintain and expand the platform after the initial MVP, working directly with the client and his current developer to sell these solutions into the Australian Market.