Engineer & technical founder · Stockholm
My path runs from Swift apps and enterprise infrastructure to data platforms, MCP-connected tools, and autonomous agent systems. Today I design lean software and AI infrastructure that teams can understand, operate, and extend.
One continuous foundation: using data to understand what is happening, analytics to decide what matters, and engineering to turn that understanding into reliable action.
During my software development studies, I was already building dashboards and using data to explain what was happening. Analytics was never a later specialization; it became the foundation I carried into every role that followed.
I founded TUVA and built native Swift and Java apps with real-time reporting, maps, notifications, and product analytics. Reaching 10,000 users in three weeks made user behaviour and feedback part of everyday product decisions.
In enterprise service operations, I began turning infrastructure and support data into repeatable pipelines, Power BI dashboards, and proactive operational insight. Reliability, ownership, and measurable outcomes became part of how I build.
As a consultant and BI developer at XLENT, I scoped and delivered Power BI dashboards across client assignments. Work with GDPR and security added governance, traceability, and responsible data use to the analytics foundation.
At Husqvarna, I used 40,000 daily sessions, 1,470 survey responses, and 30 customer visits to shape a B2B platform across 35 markets. Data connected product strategy to customer reality and operational outcomes.
I began testing ChatGPT close to its public arrival. During 2023 it moved into my daily work for analysis, writing, problem-solving, and code — first through prompts, then through a deliberate library of reusable workflows.
At Husqvarna, I moved from reporting into end-to-end data engineering: reusable ingestion from enterprise systems, 100 Databricks jobs, 200+ notebooks, and 631 governed tables — followed by forecasting, NLP, and applied machine learning.
The first coding loop was simple: ask GPT, copy, paste, test, and refine. It quickly evolved into building full sites and applications conversationally — an early version of what would later be called vibe coding.
Alongside application development, I connected models to Zapier, n8n, APIs, and scheduled processes. Prompts became reusable flows that could collect data, transform it, trigger actions, and improve through feedback.
Models became part of the actual engineering loop: reading the codebase, using context from the computer, implementing, testing, and shipping from the IDE. MCP then connected that loop to enterprise data, documentation, and tools.
I built the operating infrastructure around the models: orchestration, routing, memory, observability, scoped tool access, approvals, and failure handling. The stack now runs 10+ live projects and 33 n8n workflows.
I am developing a Company OS that connects goals, context, specialist agents, tools, memory, execution, and human oversight. MCP data pipelines give those agents controlled access to reliable data — turning a long analytics foundation into an operating system for autonomous work.
For the longer version, see about me and how I work.
Selected work
Product creation, enterprise-scale data engineering, and a self-funded agentic platform — each shown with my role, the operating scale, and what the work required.
Enterprise data & ML · 2023–2026
Solo technical lead — architecture, implementation, and operations
Built a governed Databricks platform that connects more than 20 source systems to analytics, machine learning, support intelligence, and reusable data products.
Evidence
What this demonstrates: I can own complex, long-horizon data systems from ingestion and governance to models and decision interfaces.
Read case studyAgentic infrastructure · 2023–now
Founder, architect, and operator
Designed a self-hosted operating platform where specialized agents, models, data, tools, approvals, and observability work together on real production tasks.
Evidence
What this demonstrates: I can turn emerging AI capabilities into an understandable, governed, and cost-conscious operating system.
Read case studyProduct development · 2014–2016
Founder and mobile product developer
Took a neighborhood-safety product from concept to native Swift and Java apps, a real-time backend, product analytics, and rapid early adoption.
Evidence
What this demonstrates: I have built products from zero, learned from real user behaviour, and made difficult product and business decisions.
Read case studyn8n, LiteLLM, MCP, Coolify, Hetzner — the agentic engineering stack I use to build data pipelines, MCP infrastructure, and run 10+ live projects on ~€35/month. Full details: see my toolstack.
View Full Toolstack