Chief Intelligence Architect

About Lumenfolk

Lumenfolk is building a portfolio of AI-native products designed to help people and organizations make better decisions, preserve context, and turn knowledge into usable intelligence.

Our first B2B product, Brain Console, helps organizations govern how context enters AI systems. It is not generic enterprise search, AI memory, or a simple knowledge base. It is a governance gateway for organizational context injection.

CanonIQ is the intelligence layer at the center of this vision. It is intended to help transform scattered knowledge, decisions, documents, workflows, and operating context into a more usable, trustworthy, and evolving intelligence system.

We are looking for a rare technical and strategic partner who can help design the intelligence spine of the company.

The Role

Lumenfolk is seeking a Chief Intelligence Architect to help architect, evaluate, and continuously improve the AI and knowledge systems powering CanonIQ, Brain Console, and future Lumenfolk products.

This is not a prompt engineering role. It is not a traditional ML research role. It is not a narrow software engineering role.

This is a founder-track architecture role for someone who can think deeply about LLM systems, knowledge architecture, retrieval, evaluation, governance, product intelligence, and long-term technical defensibility.

The right person will become a close thought partner to the founder and help answer one of the company's most important questions: How do we build intelligence systems that become more useful, trustworthy, and valuable over time?

What You'll Own

• AI Systems Architecture: Design the technical architecture for Lumenfolk's AI-powered products.
• Define how models, tools, agents, workflows, retrieval systems, and user interfaces work together.
• Make decisions about model selection, model routing, context windows, orchestration, cost, latency, and reliability.
• Help determine when to use RAG, long-context models, fine-tuning, structured workflows, deterministic logic, or human review.
• Create scalable architecture patterns that can support multiple Lumenfolk products.
• Knowledge Architecture: Design the knowledge structure behind CanonIQ and related products.
• Develop taxonomies, schemas, metadata standards, memory models, and source hierarchies.
• Improve retrieval precision, citation quality, and context relevance.
• Help define how knowledge enters, changes, expires, gets verified, and becomes usable by AI systems.
• Build systems that distinguish raw source material, working ideas, canonical knowledge, product runtime guidance, and user-specific context.
• LLM Evaluation and Quality: Create evaluation frameworks for AI product behavior.
• Define benchmarks for accuracy, usefulness, groundedness, retrieval quality, hallucination reduction, and task completion.
• Build regression testing processes so the system improves without breaking prior capabilities.
• Establish release-readiness standards for AI behavior.
• Translate vague product quality concerns into measurable evals.
• CanonIQ Product Intelligence: Help shape CanonIQ from concept into a durable intelligence product.
• Partner with the founder on product strategy, technical feasibility, user experience, and differentiation.
• Identify where intelligence should be automated, governed, surfaced, constrained, or handed back to humans.
• Help turn founder insight, market learning, product usage, and organizational knowledge into product architecture.
• AI Governance and Trust: Design systems that support transparency, source traceability, permissions, review gates, and appropriate human oversight.
• Help define safe boundaries for AI-generated recommendations, memory, automation, and decision support.
• Partner on trust, compliance, privacy, and explainability requirements.
• Ensure Lumenfolk's products are not just impressive, but dependable.
• Technical Leadership: Help define the early technical roadmap.
• Support hiring and evaluation of engineers, contractors, vendors, and AI implementation partners.
• Review technical work for architectural quality.
• Translate between founder vision, product requirements, and technical execution.
• Build the operating discipline needed for an AI-native company to improve over time.

What Success Looks Like

In this role, success means Lumenfolk has a clear and defensible intelligence architecture.

• A coherent technical architecture for CanonIQ.
• A measurable AI evaluation system.
• A knowledge architecture that can scale across products.
• Clear model, retrieval, memory, and governance strategies.
• A roadmap for turning CanonIQ into a trusted intelligence layer.
• Technical standards that make future products easier to build.
• A stronger bridge between founder vision and executable AI systems.

You Might Be a Fit If You Have Experience With

• LLM application architecture.
• AI systems design.
• Retrieval-Augmented Generation.
• Vector databases and embeddings.
• Knowledge graphs, ontologies, taxonomies, or semantic systems.
• Evaluation frameworks for LLM products.
• Agentic workflows and tool orchestration.
• AI product development.
• Data architecture or information architecture.
• Model selection, model routing, and AI infrastructure decisions.
• Building early-stage technical systems from ambiguity.

Ideal Background

The strongest candidates may come from roles such as Principal AI Architect, AI Systems Architect, LLM Architect, Staff or Principal Applied AI Engineer, AI Platform Architect, Knowledge Architect, Founding AI Engineer, Technical Product Architect, or CTO / technical co-founder of an AI-native product.

We care less about exact title and more about your ability to architect intelligence systems that work in the real world.

What This Role Is Not

This is not a narrow prompt engineering role, a traditional enterprise IT role, a pure academic ML research role, a generic full-stack engineering role, a chatbot builder role, or a vendor management role with no technical depth.

We need someone who can help design the core intelligence system, not simply integrate APIs.

Compensation Structure

This is an early-stage, founder-track opportunity. The company is not yet structured for a traditional market-rate executive salary from day one.

Compensation is expected to be built around a combination of meaningful equity, milestone-based compensation, possible consulting or fractional compensation during an initial trial period, and future salary as financing, revenue, or company stage supports it.

For the right person, this role may evolve into a CTO, Chief AI Officer, or technical co-founder-level seat. The final structure will depend on experience, time commitment, contribution level, risk tolerance, and whether the relationship begins as advisory, fractional, founding executive, or co-founder track.

Suggested Starting Structure

We expect to begin with a strategic architecture sprint before committing to a permanent structure.

• CanonIQ architecture audit.
• Brain Console intelligence systems review.
• 12-month AI architecture roadmap.
• Knowledge architecture proposal.
• LLM evaluation framework.
• Technical hiring and build plan.
• Recommendation on what should be built, bought, automated, governed, or deferred.

The Person We Are Looking For

We are looking for someone who can challenge, sharpen, and extend the founder's thinking.

You should be able to say: 'That is a product idea, but not yet an intelligence system.' 'This does not need more prompting. It needs evals.' 'This should be governed before it is automated.' 'The knowledge architecture is the product moat.' 'Here is how we make this measurable.' 'Here is what we should not build yet.'

The right person will bring technical depth, architectural taste, strategic judgment, and the courage to help build something ambitious from first principles.

How to Apply

Please email your resume, LinkedIn, portfolio, GitHub, writing sample, or relevant work examples to hello@lumenfolk.co with the subject line: Chief Intelligence Architect — Lumenfolk.

Include a short note answering: 1) What is the most sophisticated AI or knowledge system you have helped design? 2) How do you think about making LLM-based products more trustworthy over time? 3) What would you want to understand first before architecting CanonIQ?