I connect data & AI strategy, architecture and hands-on implementation.
Certified Chief Data Officer.
Quantitative Risk at Viridium.
I work at the intersection of data, AI, technology, governance and risk. I am interested in turning data & AI strategy into something people can actually use in their work. That takes me from enterprise architecture through to hands-on technical implementation.
At Viridium, I designed and built a corporate AI & analytics platform. It enables teams to develop their own analyses and AI applications and put them into production within shared governance, security and software engineering standards. I build the data architecture and self-service capabilities behind it, translating business and regulatory requirements into technical solutions.
My background in business informatics, quantitative finance and actuarial science helps me bring business, risk management and technology together. My focus is on data & AI strategy, platforms, data architecture and governance — and on enabling more people across an organisation to work with data and AI.
My professional background on LinkedInFor me, a data strategy becomes tangible when people know how to access data, understand it and use it in their work. I develop data architectures that bring together business context, ownership and technical access.
One starting point was my work in risk management. To see across departments, I need the raw data behind their reports. It is at this level that I can connect information and understand how it fits together.
I want other teams to have that access too. I put data mesh principles into practice and build self-service capabilities with shared guardrails. Data should be usable across departmental boundaries, while knowledge of its meaning and responsibility for it stay with the business domains.
CDO as a Service · HolonexusI designed and built an analytics platform based on Posit, R and Python, then expanded it for use across the company. It grew out of the need in risk management to process large amounts of data from different parts of the business.
Today, the aim is for teams to develop analyses independently, share results and run applications in production. Shared software engineering, security and governance standards are built into the platform. I continue to develop it and work on analyses myself, which keeps me close to the needs of the people using it.
My open-source projectsHands-on implementation also means writing software myself. I build tools for working with data and for everyday development workflows. Making them open source lets others use, understand and extend them, from R packages to extensions for Positron.
With DataRaft, I want to turn individual R scripts into data products whose results can be traced and explained. The toolkit brings together data sources, reusable processing steps, quality checks and versioned results. This makes it possible to trace a report back to the data and rules behind it. The project also includes a companion book and a hands-on course.
In development: APIs and integrations are still experimental.
Positron Deck puts everyday Positron actions on Stream Deck buttons: run code, inspect data, switch layouts or combine several steps into a workflow. I develop an IDE extension and its companion Stream Deck plugin for Windows. They support work with R, Python and Quarto, including in Posit Workbench.
jirar retrieves Jira issues and projects directly in R for analysis. It supports Jira Cloud and Jira Server/Data Center, returning data as tables that are easy to work with for reporting and automated analyses.
solvency2rfr downloads the Solvency II risk-free interest-rate curves published by EIOPA and prepares them as R tables. The monthly data can then be used directly in actuarial models and reproducible analyses.
I build platforms that let teams use AI in their day-to-day work and develop their own applications. Using Azure AI Foundry, I created the technical foundations for using AI across the company.
In my risk management role, I am also responsible for the AI policy. For me, architecture and governance belong together. I translate AI, information security and regulatory requirements into technical solutions that help colleagues work within shared guardrails.
People often choose the easiest path. That is why I pair rules with practical tools and infrastructure. I want to make the easiest path the intended one, so that AI can become a natural part of everyday work.
I have worked in Risk Analytics at Viridium Group since 2018. I connect business, technology and mathematics to understand risk in the context of the whole organisation. Bringing data and business perspectives together across silos helps me support risk assessments and decisions.
This perspective across the business also led to my work on data architecture, analytics and AI platforms. I want to build what I need for my own analyses in a way that other teams can use too. That includes making governance requirements part of the technical implementation.
That breadth is what I enjoy. I can learn about many different subjects, understand how they affect each other and see how they fit into the bigger picture.
I specialised in insurance during my studies. Afterwards, I wanted to put that knowledge into practice and combine it with my enthusiasm for data, technology and mathematics.
I see insurance as a social institution that can give people security and protect their standard of living. Many people share a risk so that no one has to face its consequences alone. To me, this pooling of risk is a deeply social idea. I want to help make insurance better and more humane.
Through Holonexus, I want to share the experience I have gained over recent years with other organisations, as Chief Data Officer as a Service. That includes developing data & AI strategy, building data architectures and platforms, and making governance work in everyday practice.
I want to share what has worked in my own practice and explore together what fits each organisation. I would like to support the journey from strategy through architecture to practical implementation, creating capabilities that people can continue to develop independently.
holonexus.deWhen do investments in data and AI actually create enterprise value? I am interested in how technical and organisational capabilities interact, the order in which they develop and how organisations learn from new evidence. This also raises the question of when to continue an experiment and when to stop.
Working Papers
Learning, Remediation, and the Option to Stop
Should an organisation improve its foundations first or learn by trying new tools? I examine how diagnosis, pilot costs and the option to stop affect transformation sequencing. In the theoretical model, starting with tools is preferable only within a specific cost range, rather than as a general rule.
Evidence Queues, Recoverability, and Priority Rights
A worthwhile experiment can still be privately unattractive when evidence is correlated and time or priority rights expire. I examine how incentives can support both the start and later continuation of research. The theoretical model derives implementation conditions and funding limits within a restricted experimental setting.
Six Interdependent Capabilities and the Dynamics of Transformation
In this theoretical model, I examine six capabilities: People, Data Quality, Data Architecture, Data Governance, Automation and AI. Their interaction determines whether scaling creates value or amplifies errors. The paper describes when reliability should be improved first and how shared investment can enable transformation. Synthetic examples illustrate these relationships.
Data become more valuable when others can use them. At the same time, their use requires reliable foundations and appropriate incentives. I examine how data sharing, provenance and independent challenge interact, including when AI is used.
Working Papers
Admission, Producer Incentives, and the Financing of Rival Support
Data can be reused widely, but the support provided by their producers remains scarce. In a theoretical model, I examine how internal accounting can encourage valuable data use alongside investment in quality and documentation. For the accounting rules studied, I identify the resulting financing and participation constraints.
Independent challenge, fabricated opposition, and coordination
A dissenting report does not establish that information was produced independently. I examine how payments, provenance audits and the use of reports interact when information production is hidden. Within the contract designs studied, the theoretical model shows how coordination needs change the appropriate audit and information architecture.
Provenance Before Verification: Timing, Repair Coverage, and Adoption
Recognizing that an AI output is wrong does not reveal which reusable data source caused the error. I examine how the timing of provenance information affects investment in data quality, verification and AI. Under the theoretical model’s conditions, early provenance can complement these investments and make AI adoption worthwhile.
Which information is sought, shared and used in decisions? I am interested in how incentives, measurability and scarce review capacity shape that selection. My models examine the design of rules and decision rights.
Working Papers
Costly Verification and the Audit of Unchanged Decisions
I examine how career and reputational concerns affect the search for evidence that could challenge an earlier decision. In the theoretical model, the distortion lies less in ignoring acquired evidence than in failing to commission costly checks. Targeted audits of unchanged decisions can improve that incentive; their value depends on audit costs.
Scarcity, endogenous applications, and the optimal direction of evidence adjustment
Evidence requirements shape both which innovation projects receive funding and which teams apply in the first place. I examine how differences in measurability interact with scarce budgets. In the theoretical model, whether harder-to-measure projects should face weaker or stronger requirements can reverse solely because the available budget changes.
Information Access, Strategic Proposals, and Scarce Review
Who should receive scarce review capacity when managers decide whether to submit a proposal? I compare preassigned reviews with allocation contingent on submitted proposals. The theoretical model shows that achieving a target at low cost does not guarantee that participants will coordinate on the intended equilibrium.
Rational Inference, Factual Verification, and Metric Governance
Accurate metrics can still present a selective picture when the most favourable ones are chosen. In a theoretical model, I examine how such reports affect internal investment decisions and what random audits can achieve. I distinguish factual verification from governance over which metrics are used.
Feasible Auditing, Report Use, and the Nonmonotone Demand for Protection
How should organisations combine independent audits with the use of employee reports for performance evaluation? In the model, I distinguish prevention, discovery and disclosure. Under the rules studied, a higher value of information does not automatically call for more protection for reporters; the appropriate combination of auditing and report use can change direction.
I am among the early researchers involved in the modern rediscovery of tontines. During my PhD, I revisited this long-overlooked concept and explored its role in retirement provision today.
Tontines date back to the 17th century and are among the historical predecessors of modern life annuities. Their central idea is collective provision: the assets of members who die benefit those who survive.
Using actuarially fair models, I studied how tontines could be used in modern retirement provision. My research covered the sharing of longevity risk, combinations with other retirement products and the fair valuation of an early exit.
In my view, tontines have the potential to offer a superior alternative to conventional retirement products when their design fits individual needs. I examined these questions in depth academically, resulting in my doctoral thesis Essays on Tontine Annuities and several publications.
Working paper & publications
A Sharp Frontier for Mortality Pooling, Model Error and Death Benefits
How fair does an open tontine remain when its members’ mortality is misestimated? I examine the relationship between mortality pooling, model error and death benefits. Within the model class studied, I derive a sharp boundary between these objectives and analyse voluntary exits. Reproducible checks and R simulations complement the mathematical results.
An Innovative Instrument for Longevity Risk Management in an Ageing Society
under systematic longevity risk
In: The Past, Present, and Future of Tontines
My academic background is in quantitative finance and actuarial science. During my PhD, I studied models for retirement provision, particularly tontines. This analytical foundation continues to shape my work across data, technology and risk.
During my PhD, I presented my research at numerous conferences and discovered how much I enjoy speaking about my work. I continued throughout my professional career. Today, my talks, panels and podcasts focus on making data and analytics useful across an organisation, building platforms and putting the transition to a data-driven organisation into practice.
What I value most is the conversation that follows. A talk lets me share my experience with a wider audience; the discussion afterwards brings new questions and perspectives. I often learn the most from those conversations myself. That is why I enjoy it so much.
Data Driven Insurance · Versicherungsforen Leipzig
Leipzig, Germany
CIONET Online Community Event
Online
Data Driven Insurance · Versicherungsforen Leipzig
Leipzig, Germany
TDWI München · Data Strategy & Data Governance
Munich, Germany
Data Driven Insurance · Versicherungsforen Leipzig
Leipzig, Germany
Munich, Germany
American Economic Association
San Diego, USA
HEC Lausanne
Lausanne, Switzerland
ARIA Annual Meeting · American Risk and Insurance Association
San Francisco, USA
Neu-Isenburg, Germany
Deutsche Aktuarvereinigung · Deutsche Gesellschaft für Versicherungs- und Finanzmathematik
Düsseldorf, Germany
Jahrestagung des Deutschen Vereins für Versicherungswissenschaft
Berlin, Germany
Amsterdam, Netherlands
45th EGRIE Seminar · European Group of Risk and Insurance Economists
Nuremberg, Germany
ARIA Annual Meeting · American Risk and Insurance Association
Chicago, USA
St. John’s University
New York City, USA
Deutsche Gesellschaft für Finanzwirtschaft (DGF)
Ulm, Germany
ARIA Annual Meeting · American Risk and Insurance Association
Toronto, Canada
Jahrestagung des Deutschen Vereins für Versicherungswissenschaft
Berlin, Germany
CEAR / MRIC Behavioral Insurance Workshop
Munich, Germany
Augsburg, Germany
European Group of Risk and Insurance Economists
Limassol, Cyprus
Jahrestagung des Deutschen Vereins für Versicherungswissenschaft
Vienna, Austria
Together with two co-authors, I have written two companion books. They grew out of my professional experience: using data well can make a real difference to an organisation. But becoming data-driven is a sociotechnical challenge. Technology, people and the organisation need to work together across many different areas.
In much of the literature, we found a gap between the ideal being described and its practical implementation. Individual aspects receive detailed attention, but the connections between them often remain unexplored. We developed a holistic framework that brings these areas together. The non-fiction book describes the approach; the companion novel lets readers experience it through a story.
I have been a huge Eintracht Frankfurt fan since childhood and a season-ticket holder for many years. Whenever I can, you will find me in the Nordwestkurve or travelling to away games, in Germany and abroad.
Every winter, I head out on my snowboard. France is my favourite destination, but I also ride in Austria, Italy and Germany. Snowboarding is my favourite kind of holiday, even though I am otherwise very much a summer person.
Every year, I spend several weeks at a time travelling in my self-converted T6, exploring Scandinavia and southern Europe. One three-month trip took me all the way to Albania.
During the conversion, I learned about electrics, solar power, woodworking and insulation. Learning something new and trying it out myself is what I enjoy most about it.
I enjoy walking and taking the time to look closely. On foot, I notice things I would miss driving past. A camera adds another perspective: a different position or composition can completely change how a scene feels.
What interests me most in photography is the process of getting to the final image. I have an idea of how I want a photograph to look, then experiment with how to achieve it. I keep learning and putting myself in new situations. It adds another dimension to being out and about.
I shoot with a Lumix S9. My favourite lenses are 75 mm f/2 and 10 mm f/2.5.
Loading photographs…
You can reach me at mail@janweinert.com or use this form.
