
01_
status = 'Datamatiker · 3rd semester',
passion = 'Software, AI tooling & Product',
location = 'Aalborg, Denmark',
age = 22,
languages = {'Danish', 'English'},
discipline = 'Disciplined by iron, driven by code.'
// use keyboard
// arrows to play
// food left

const jakob_2010 = {
passion: 'football',
skills: '[ kicking things ]',
}
// ↓ 16 years of character development later
const jakob_now = {
passion: 'code',
skills: '[ actually useful things ]'
}
Building the AI tooling a product team runs on.
From September 2026 I sit in the Product Management team at EG Digital Welfare — welfare software for Danish and Norwegian municipalities. Not product features: the agents, Claude skills, evals and data pipelines behind how the team works.
Agents & Claude skills
Research, documentation and reporting agents the team uses day to day — built and maintained in GitHub.
Evals
Test sets and metrics that decide whether an agent or workflow is good enough to ship — and catch regressions.
Product data
Pipelines from product and customer data to analyses, insights and visualizations for PMs and leadership.
Enablement
Training Product Managers, UX designers and Business Analysts in the tools — and sparring on new AI approaches.
UCN Aalborg
Computer Science (Datamatiker) at UCN — Java, SQL, system development and software architecture, with C#/.NET this semester. Alongside the study I ship my own products and work as Student Assistant in Product Management at EG.
TypeScript / Node
Main language for everything I ship today: CallGuard, the analytics platform, Studie3 and this site. Node.js, Express, Fastify, REST APIs, Vitest.
SQL / Azure / Postgres
SQL Server on Azure, PostgreSQL via Supabase. Relational schemas, validation at ingest, views and the queries behind CoachData's live analytics and the event platform.
Claude Code / AI tooling
Claude Code as a daily tool and as infrastructure: custom skills, a scheduled agent pipeline in Studie3, a Gemini analysis flow in CoachData — and now agents and evals for the PM team at EG.
React / Next.js
React + Vite for CoachData, Next.js 15/16 for CallGuard, Studie3 and this portfolio. Deployed on Vercel and Railway.
LuaU
Since 2017. Four published games with 230K+ combined visits — combat systems, progression, GUI, monetization and the SDK for my own analytics platform.
Java
Core academic language. Inventory, sales and reservation systems with UI, controllers, domain model and database access — built in teams with UML and Git.
Production_Builds

CallGuard
Lead-recovery SaaS that turns missed service enquiries into instant replies, bookings, follow-ups and a visible sales pipeline.
Open website
CoachData
Handball analytics SaaS. Shot maps, player ratings, phase analysis, assist networks — built for real clubs.
Studie3
Study app with a Claude agent in the loop: answers written in the app come back as feedback, notes and flashcards — automatically.
Read moreRoblox Analytics
Event platform prototype: LuaU SDK → Fastify API → Azure SQL → Next.js dashboard with retention, funnels and monetization.
Read moreLive games
Four published games, operated end-to-end. 230K+ visits and 9K+ favourites — live numbers on the projects page.
Read more