# Invenire

> Founder-led AI engineering studio. We help teams turn AI prototypes
> into production systems — agents, retrieval, vision, evals, telemetry,
> and the engineering glue. Founded and run by Arve Sølland.

Frontier AI for the part after the demo. Invenire is a founder-led AI
engineering studio based on the Gold Coast, Australia. I design and ship
agentic systems, retrieval and vision pipelines, evaluation
infrastructure, analytics products, and AI-native engineering workflows
for teams that need AI to work outside the demo. The flagship product is
Analytics Agent, a multi-agent analytics app for Shopify merchants.

The studio takes a small number of engagements at any one time —
typically one or two — across commerce, climate / geospatial / regulated
work, founder-led builds, and engineering teams going AI-native. The bar
is whether the work has to actually run in production.

## Pages

- [Home](https://invenire.com.au/): Studio overview, services summary, flagship product, selected work, founder bio, contact.
- [Services](https://invenire.com.au/services/): Detailed write-ups of the four service offerings (AI strategy and architecture, production AI systems, evals and instrumentation, AI-native engineering practices) with engagement shapes and "good fit when" callouts.
- [Selected Work](https://invenire.com.au/work/): Five production case studies, each with a deep page — Analytics Agent (Shopify), Plant (autonomous engineering platform), EcoCommons (national modelling platform), an autonomous model-training agent (confidential), and an image annotation and QA platform (confidential).
- [About](https://invenire.com.au/about/): Founder background and career arc (frontier AI work, EcoCommons, Pacific iCLIM), operating principles, tech stack, education, references, contact channels.
- [Contact](https://invenire.com.au/contact/): Discuss a production AI engagement. Cal.com booking, email, LinkedIn, "what to include in your first message", fit filter, FAQ, response-time commitment.

## Services

- [01 / AI strategy and architecture](https://invenire.com.au/services/#ai-strategy): What to build, what to buy, what to ignore. Architecture review, opportunity mapping, model and tool selection, build/buy/partner recommendation, and a phased implementation plan.
- [02 / Production AI systems](https://invenire.com.au/services/#production-ai-systems): Agents, retrieval, vision, multi-modal workflows designed to run beyond the demo — with observability, fallback paths, permissions, review loops, and evaluation built in from the start.
- [03 / Evals and instrumentation](https://invenire.com.au/services/#evals-and-instrumentation): LLM-as-judge harnesses, golden datasets, PR regression gates, traces, and cost / latency / drift telemetry. Quality as a number you can fail CI on.
- [04 / AI-native engineering practices](https://invenire.com.au/services/#ai-native-engineering): Adoption of Cursor, Claude Code, agentic workflows, eval-first development, and review loops as a team discipline.
- [Engagements](https://invenire.com.au/services/#engagement): Five shapes — diagnostic and architecture sprint, production AI build, evals and observability review, AI-native engineering enablement, and principal advisory. Scope and investment are discussed after there is enough context to be useful.

## Selected work

- [Analytics Agent](https://analytics-agent.app/): Six specialist AI agents coordinated by a router. Reads a Shopify merchant's GA4, Search Console and store data daily and delivers a Mission Brief — three to five plain-English findings. Fifteen-minute anomaly detection, GA4 audit with one-click fixes. Live in the Shopify App Store.
- [Plant](https://invenire.com.au/work/plant/): A self-hosting platform of seven services in five languages that lets one operator run like a full engineering org — externalised state, objective completion verification, agent-first tooling, a loop that builds itself. plnt.dev.
- [EcoCommons](https://invenire.com.au/work/ecocommons/): Australia's national platform for ecological and environmental modelling. Technical Lead & Solutions Architect at Griffith University; nine partner institutions, ~$5M over three years, reproducible-by-design. Public on ARDC.
- [Autonomous model-training agent](https://invenire.com.au/work/autonomous-training/): Confidential client. An autonomous agent that does the ML research engineer's job — hypothesis, training code, experiment, score, repeat — for hundreds of cycles, with the engineering in the anti-gaming guardrails. Reference under NDA.
- [Image annotation & QA platform](https://invenire.com.au/work/annotation-platform/): Confidential client. Turns many fallible reviewers into one trustworthy, audit-grade, model-ready dataset — weighted consensus, surfaced disagreement, full provenance. Reference under NDA.

## Founder

- [Arve Sølland](https://invenire.com.au/about/): Principal AI Engineer and founder of Invenire. Twenty years shipping production software, from national-scale research infrastructure to modern agentic AI systems. Before founding Invenire, Technical Lead and Solutions Architect on EcoCommons and Technical Lead on the Pacific iCLIM project (Griffith University / SPREP). Based on the Gold Coast, Australia.

## Contact

- Email: hello@invenire.com.au (primary, studio inbox)
- LinkedIn: https://linkedin.com/in/arvesolland
- Founder-led. You work directly with Arve.
- Based: Gold Coast, Australia (AU · GMT+10)
- Response time: within 48 hours, Australian business days
- Engagement size: typically 1–2 clients at a time, scoped after context.

## Operating principles

- **Currency** — The AI frontier changes quickly. I track the models, tools, patterns, and failure modes closely, then apply the parts that hold up under production constraints.
- **Evidence** — AI systems need measurement. I prefer evals, traces, golden datasets, regression gates, cost curves, and failure taxonomies over opinions about whether a model "seems better."
- **Reliability** — Non-determinism is not an excuse. Production AI needs bounded responsibilities, structured outputs, observability, fallback paths, human review where appropriate, and tested recovery.
- **Focus** — The goal is not to use the most AI. The goal is to ship the smallest system that creates leverage and can be trusted.

## Engagements

- Five shapes: diagnostic and architecture sprint, production AI build, evals and observability review, AI-native engineering enablement, and principal advisory.
- Scoped after enough context to be useful. No public day rate or fixed-price bands.

## Optional

- [Full site as markdown](https://invenire.com.au/llms-full.txt): Every page concatenated as a single markdown document, suitable for ingesting the whole site in one fetch.
- [analytics-agent.app](https://analytics-agent.app/): Standalone product site for the flagship — separate brand, same studio behind it.