Taking on new work

Everyone's talking about AI. Nobody's shipping it.

I build AI systems that actually work in production - not demos. Strategy to deployment in regulated industries where the stakes are real.

Agentic Systems in Production
13+
Years at the Forefront of Tech
1,000+
Users Scaled
100%
Client Retention
50+
Workflows Automated

Services

I solve problems others walk past

Focused on the gap between AI strategy and production reality - where most engagements stall.

Too many tools, no clear path

You're inundated with AI tools and vendors, each promising to transform your business. The result? Analysis paralysis and zero shipping velocity.

Buried in manual work

Your team spends hours on tasks that should be automated. The fix exists - but no one has the time or depth to build it properly.

High stakes, no room for error

You operate in regulated industries where mistakes mean compliance failures or worse. Generic AI tools weren't built for your reality.

No one to own it

You've got consultants who advise and developers who build - but no one who does both and stays accountable end to end.

Process

From first call to production in weeks

Three engagement phases. One accountable person. No hand-offs, no gaps.

  1. 01
    Discovery 1–2 weeks

    Understand your current state and map the opportunity

    We audit your workflows, data, and tooling to identify where AI creates the most leverage. You get a prioritised roadmap - quick wins mapped against strategic plays.

    • Current-state workflow audit
    • Automation opportunity map
    • Prioritised roadmap (quick wins vs strategic plays)
    • Technical feasibility notes
  2. 02
    Build 4–8 weeks

    Ship a production-grade AI system, not a demo

    I design, build, and deploy the system - integrated into your existing workflow and tested with real users. No hand-off to a separate engineering team; I own it end to end.

    • Working production system
    • Integration with existing tools and data
    • User acceptance testing
    • Documentation and runbooks
  3. 03
    Maintain Ongoing

    Monitor, iterate, and expand as you grow

    AI systems need human oversight. I stay on as a fractional partner - monitoring quality, iterating on the model, and expanding capability as new opportunities emerge.

    • Weekly quality monitoring
    • Model iteration and fine-tuning
    • Incident response
    • Quarterly expansion planning

Work

Selected case studies

View all work
AI Multi-Agent Pharma GCP

Multi-Agent FDA Document Review

Clinical-stage biotech company · Biotechnology / Pharmaceutical · October 2024 - February 2025

Designed and deployed a multi-agent AI system for FDA regulatory document review. 6 specialized agents process 100+ page documents at ~18 seconds/page with 15-40% duplicate removal. Multi-million USD savings projected.

  • 60-70% reduction in manual review time
  • ~18 seconds per page processing
  • Multi-million USD savings (client-validated)
Google Cloud Platform Vertex AI (Gemini 1.5) LangChain Terraform
Data Governance Pharma Enterprise Compliance

Enterprise Data Governance Transformation

Specialty pharmaceutical company · Pharmaceutical · April 2025 - September 2025

Led technical delivery of a €400K data governance transformation for a specialty pharmaceutical company across 13 countries - coordinating 25+ stakeholders, mapping 7 data domains, and delivering a framework for regulatory compliance.

IFS ERP Veeva Vault Power BI Data governance frameworks
Data Migration Pharma Risk Management Veeva

When to Halt a Migration

mRNA biotech company · Biotechnology / Pharmaceutical · April 2024 - October 2024

Led assessment of a 400,000+ document clinical trial migration. Through rigorous data profiling, reduced scope by 92% - then recommended halting when validation revealed critical data quality issues.

Veeva Vault eTMF ML-assisted document classification Content hashing / similarity scoring
PLG SaaS Startup Growth

Product-Led Growth Engine

Norwegian clinical research startup · Healthcare SaaS · April 2020 - February 2023

Joined as employee #5 at a pre-revenue SaaS startup and built the complete commercial infrastructure - PLG strategy, 8-system tech stack, customer success, and enterprise readiness. Scaled 0 → 1,000+ users with 100% retention.

Segment CDP HubSpot Chargebee/Stripe Amplitude Freshdesk
Data Engineering SEO BigQuery GCP

GSC → BigQuery Pipeline

Copenhagen-based SEO agency · Digital Marketing · September 2025 - Present

Built an automated data pipeline ingesting Google Search Console data into BigQuery for a 60-customer SEO agency. 53M+ rows, 70,000+ daily ingestion, solving GSC's 16-month data retention limit.

Google Cloud Run BigQuery Cloud Scheduler Google Search Console API

Let's work together

Ready to ship something real?

I take on a small number of engagements at a time so I can stay deeply involved. If the timing works, let's talk.