Direction
AI strategy
Start with the business problem. Identify where AI is worth using, what it could change, and what should be left alone.
AI consultant · founder · forward deployed engineering
AI consulting from strategy to production
I work with companies to identify where AI is worth using, assess models and software, understand data and technical constraints, and build or integrate the right solution.
What I do
Strategy, technology choices and implementation
Sometimes the best answer is software you can buy. Sometimes it is an agent, a custom workflow, or a model that needs to be trained for a specific problem. I help companies make that call and, when needed, implement it.
Technology
Models and software
Compare commercial and open models, existing products, RAG, agents, fine-tuning and custom ML without assuming one approach is best.
Execution
Build and integrate
Prototype and ship agents, internal tools, automation and integrations around the systems and workflows already in place.
Adoption
Make it stick
Test with the people who will use it, understand the edge cases, and get it reliable enough for daily work.
Buy, integrate, build or train
I do not start with a preferred technology. The right answer may be an existing SaaS product, an agent connected to current systems, a custom application, or a model trained for a narrow problem.
BuyUse the best existing software.
IntegrateConnect AI to current systems.
BuildCreate the missing application or workflow.
Own modelsTrain or adapt models when it genuinely pays off.
In practice
What the work actually looks like
Forward deployed work means staying close to the company and owning the path from the first question to a system running in production.
Intake and goals
Understand business goals, workflows, stakeholders and what success should look like.
Current state
Map existing software, data, interfaces, technical debt and how the work is done today.
State of the art
Review available models, products and approaches, then compare them against the actual requirements.
Requirements and limits
Clarify data quality, security, privacy, cost, latency, operations and organizational constraints.
Prototype and evaluate
Build a useful first version, define tests and evals, and validate it with real users and real data.
Roll out and improve
Integrate, productionize, measure adoption and keep learning from feedback and real workflows.
XSUPRA GmbH
I have made these trade-offs as a founder
XSUPRA GmbH was the precision-agriculture startup I co-founded and led as CEO. We built an AI product from scratch and sold it to farms across Europe.
I worked across product, customers, architecture and company strategy, and built the full AI product. That included Alora, our AI agent, retrieval and agentic workflows, predictive and soil-intelligence models, and satellite-based features.
It made me skeptical of demos that ignore the rest of the system. Data quality, integration, reliability, usability and commercial value matter just as much as the model.
100+farms across European markets
AloraAI agent built into the product
0→1from idea to paying customers
Product + companycustomers, architecture, GTM and company building
Background
Technology, business and the space in between
AI founder
Built an AI product from idea to adoption while working directly with customers, product, architecture, fundraising and company strategy.
Enterprise advisor / Solution Architect
Worked with senior and C-level stakeholders on solution design, digital transformation, cloud, integrations, security, scalability and performance.
AI consultant / Software engineer
Advised companies, artists and startups while building software, custom AI agents, assistant workflows, apps and prototypes.
Applied researcher
Developed a disaster-management information system from research concept through prototype and government-funded deployment.
Personally
Technology has never been just a job to me
I’m an entrepreneur and technologist. My head is usually full of ideas, and I have always felt the need to explore them, test them, and turn the promising ones into something real.
When I first started coding, I realised how much leverage software gives you. That never really left. I still experiment constantly, build apps for fun, and follow everything from hardware to new AI systems.
I am excited by what AI can do, but I also care about trust, privacy, sustainability, responsibility, and its effect on people and society.
Many of my best ideas come from conversations with people who see the world differently. Good products often start there.
Co-founding
I am always open to the right co-founder conversation
Especially when a hard business problem and a new technical possibility genuinely line up.
Contact
Working on an AI problem or a new company?
Tell me what you are trying to change. I am interested in consulting, forward deployed work, selected product projects, and co-founder conversations.