AI-Native Product & Systems Builder | Stockholm, Sweden / Remote
I help turn expensive, manual or unreliable workflows into practical AI-assisted systems that are easier to operate, test and trust.
My role is to frame the business problem, set the constraints, direct AI and software execution, verify the result and decide what is ready to use.
Start small: AI Workflow Audit. One workflow, one real constraint, one prioritized improvement path.
- AI Workflow Audits: find where time, cost, manual work or reliability leaks, then identify the smallest useful fix.
- AI automation and internal systems: turn repeatable operational work into practical automations, integrations and agent-assisted workflows.
- Verification and reliability: add tests, evals, readback and explicit failure states when "probably worked" is not good enough.
- Completed 5-star client work in local SEO and website optimization.
- Completed paid advisory work focused on Google Page Speed and LCP performance.
I use paid work as commercial proof and the public repositories below as inspectable technical proof.
A runnable reference for verifying what actually happened before an automated system retries or takes another action.
Shows: state verification · reconciliation before retry · explicit uncertainty
A bounded decision engine that turns agreement rules and meeting evidence into BILLABLE, NON_BILLABLE or REVIEW.
Live product · Public reference
Shows: commercial rule modeling · evidence traceability · review states
A release-verification reference that binds evidence to the exact artifact and environment being judged.
Shows: artifact identity · evidence binding · explicit INCONCLUSIVE
- Agent Forecast Foundry: verifies evidence, output structure, provider state, cost and latency before accepting an AI run.
- PriceBriefs: rejects weak competitive-price comparisons until identity, currency, availability, source quality and freshness are strong enough.
- Find the real operational or commercial constraint.
- Define the evidence, acceptance criteria and failure boundaries.
- Direct AI agents and software tools to implement the smallest useful system.
- Test the result against the state that actually matters.
- Iterate, reject or ship based on evidence.
AI gives me implementation leverage. I am accountable for problem framing, system direction, orchestration, verification and final judgment.
Before my current AI work, I spent years in investigation and evidence-heavy operational roles, with earlier studies in IT forensics and information security. That background shaped how I think about uncertainty, evidence quality and system failure boundaries.
From 2015 to 2020, I also built and operated automated FX trading systems in live markets, where software behavior had direct financial consequences. A historical third-party performance record is available on Myfxbook.
AI agents and coding tools · TypeScript / Node.js · Python / FastAPI · React / Next.js · Postgres / Supabase · OpenAI APIs · GitHub Actions
If you have one workflow where cost, manual work or reliability is becoming a problem, send me the workflow.