AI Consultant | AI Delivery & Team Enablement

Making AI work in your codebase, not in a demo.

I build AI products, and I help teams build their own. Founded and built an AI platform solo, now in invite-only beta — one adapter over many inference providers, token-accurate cost accounting, retrieval and orchestration built in dependency order — and ran AI enablement as a consultant for both engineering teams and IT leadership. My work is the unglamorous half: provider choices you can reverse, cost that survives scale, the failure modes that stay silent by default, and the governance questions that arrive with EU deployment.

[ Multi-Provider AI Platform ][ Workshops: Teams & Leadership ][ Agentic Delivery ]

High-Impact Case Studies & Roles

2026 – Present

Sushibelt — AI Workflow & Game-Asset Platform

Sushibelt

// OVERVIEW

–A working answer to what a serious AI product costs to build and run

–Providers behind one adapter, so a workflow routes each node to whatever model suits it and falls back when one fails; reserve-at-dispatch credit accounting, so every call is priced before it runs; an evaluation harness to compare providers on identical input

–The AI layer proper was built in dependency order — retrieval with pgvector before memory, memory before tools, a typed tool registry with an audit table before any autonomous orchestration, and only then a goal-driven loop bounded by a credit budget, with MCP exposure for headless use

–The ordering is the opinion: each one is only safe once the one beneath it is real

// IMPACT METRIC

In invite-only beta testing with real users since late August 2026

The AI layer — pgvector retrieval, assistant memory, a typed tool registry and an MCP server — is built in that dependency order, and goal-driven orchestration on top of them is in testing

// TECH STACK

Provider StrategyLLM OrchestrationRetrieval & MemoryInference CostEval Harness

June 2024 – April 2025

CTO / Head of Product Development

Certainly // Copenhagen, Denmark

// IMPACT & IMPLEMENTATION

–CTO of an AI SaaS in the e-commerce conversation space

–I owned the AI direction for both the platform and the organisation: where agent orchestration and LLM workflows genuinely replaced hand-built logic and where they did not, and what had to hold before a model went near a customer conversation

–I used AI internally as well, to accelerate delivery against a modernization roadmap

–And I carried it to the company's partners, explaining in their terms how the AI worked and what it was accountable for

// IMPACT METRIC

Acted as the technical guarantor for the company's partners

// TECH STACK

AI AdoptionAgent OrchestrationTechnical Guarantor

January 2024 – May 2024

Principal Consultant

Knowit // Copenhagen, Denmark

// IMPACT & IMPLEMENTATION

–Ran AI enablement inside a consultancy serving traditional enterprise clients: workshops for the internal technical team, and a separate session pitched at IT leadership

–Two audiences, one subject, two vocabularies — the technical group wanted model choice, integration patterns and where the hype stopped; leadership wanted data protection, data residency, and a clear read on what they were committing to

–That split is the core of the consulting work: the same honest answer, delivered so each room can act on it

// TECH STACK

AI WorkshopsData ResidencyEnterprise Enablement

Professional Journey

Solo Founder

Sushibelt

Founded Sushibelt in January 2026 and have built and run it alone since: an AI workflow platform for making game assets, in invite-only beta testing with real users since late August 2026.

January 2026 – Present

Product Engineer & Technical Advisor

Freelance (Eag)

Operating through eag.dk, I build customer-facing products and give technical advice to startups and early-stage companies. Since 2011, I have worked on 10 projects, delivering tailored technical solutions that bridge the gap between initial concept and market deployment — and since 2023, increasingly helping non-technical founders make informed technical decisions in early planning and scoping.

2011 – Present

CTO / Head of Product Development

Certainly

Led the technical vision, architecture, and execution of SaaS product development. Hired as lead developer to help refactor and modernize the platform, I moved into the CTO/Head of Product role to stabilize the platform, lead the team, and act as the technical guarantor for the company's partners.

June 2024 – April 2025

Principal Consultant

Knowit

Worked with traditional enterprise clients, and ran AI workshops for Knowit's internal technical team plus a separate one for IT leadership.

January 2024 – May 2024

Founder in Residence

Antler

August 2023 – November 2023

Technical Team Lead / Project Manager

ComplyCloud

Hired as Lead Developer to architect version 2 of the ComplyCloud platform, supporting a team growing from 2 developers to 30. I constructed a modern architecture that provided tools for developers, content managers, and lawyers to engage with the platform more effectively.

February 2020 – January 2023

Frontend Developer

Reto Moto

Hired as a freelancer to build up and support a modern games platform. Focused on building a proper architecture for it, and improved performance and stability dramatically for 10,000+ daily active users.

2017 – 2018

Programmer

BetaDwarf Entertainment

Early-stage developer: helped the team build out concepts and core gameplay features for the start-up's debut title.

2011 – 2012

Education

Master's Degree in Medialogy

Aalborg University

2007 – 2016

Full Portfolio

View All Projects & Sub-projects

Skill Matrix

AI Engineering & Adoption

Agentic Development Practice

Provider Evaluation & Benchmarking

Inference Cost & Token Economics

LLM Failure-Mode Debugging

AI Adoption Workshops

AI Build-vs-Buy Scoping

Data Protection & Residency

AI Governance (EU AI Act)

Architecture & Strategy

AI-first Development

SaaS Architecture

AI/LLM Integration

Technical Roadmapping

Multi-Tenant Design

Workflow Automation

AI Agent Orchestration

LLM Tooling & MCP

Retrieval (RAG / pgvector)

Multi-Provider AI Integration

Geospatial Systems

Investment Readiness

Development Stack

React

Next.js

TypeScript

Node.js

Unity

React Native (Expo)

SQL

AR Integration

DevOps & Cloud

Supabase (Postgres/Auth)

AWS Amplify

Azure

Docker

CI/CD Pipelines

Row-Level Security (RLS)

Leadership & Delivery

Team Scaling (2-30)

Planning & Scoping

Cost Allocation

Product-Market Fit

Rating

Aware

Familiar

Competent

Proficient

Expert

Adding AI to something real?

Provider choice, cost at scale, and the failure modes that stay quiet — worked through on your codebase, with your team in the room.

Start an AI engagement