VinkeRE — pricing-to-capital software for non-life reinsurance. One high-performance Python core, delivered through whatever interface fits your team — Excel, API or web.
We built a reinsurer’s full pricing infrastructure in months — and ran it in production for eight years. Now we’re building the next generation of actuarial software.
The complete modelled-risk workflow — vendor cat outputs (RMS, AIR), cyber model outputs, or parametric frequency-severity simulation. ELT/YELT ingestion → unique hashing → stochastic simulation → treaty pricing & structuring → documentation → portfolio accumulation, with roll-back by layer, peril, country and model. One common language: ELT and YELT — delivered through the interface that fits your team.
The same Python core, calibrated to your modelling framework — SST, Solvency II or internal model. Not just a tailored interface: your assumptions, your risk view, your capital structure, built into the engine.
Bespoke pricing environments and hands-on support — tools, databases, reporting — built and run in production for eight years at Toa Re Europe. The experience VinkeRE was born from.
Senior actuarial advisory delivered on our own engine — structuring reviews, portfolio studies, second opinions. Fast, because the platform does the heavy lifting; rigorous, because an actuary drives it.
A good tool is only half the story — teams have to master it. Training programmes built around the platform and your workflows, from operator onboarding to underwriter-level output literacy.
Shown here: VinkeRE Modelled Risks on a cat programme. The same core powers VinkeRE Non-Life.
VinkeRE runs the whole chain — vendor model outputs in, capital-aware decisions out. ELTs are forged once into seeded, reproducible YELTs; treaties are priced and structured with full drill-down by layer, event, peril and vendor model; portfolios accumulate gross and net of retrocession; capital is allocated at the point of underwriting.
The Forge — where vendor model outputs become pricing-ready. Every ELT is fingerprinted once, archived in Parquet, and simulated into YELTs with proper secondary uncertainty via pre-simulated Year-Event Quantile Tables (YEQT) — seeded, stored, reproducible to the row.
Simulate once, price everywhere: no duplicate runs, no drift between analyses, a full audit trail by design.
| Layer | Attach | Limit | EL % | Tech. Premium | RoL |
|---|---|---|---|---|---|
| Layer 1 | 10 | 15 | 9.8%→11.0%+12% | € 3.60M→€ 4.00M | 26.7% |
| Layer 2 | 25 | 25 | 5.6%→6.6%+17% | € 4.20M→€ 4.85M | 19.4% |
| Layer 3 | 50 | 25 | 3.2%→3.9%+22% | € 2.80M→€ 3.35M | 13.4% |
| Layer 4 | 75 | 50 | 1.6%→2.1%+30% | € 3.20M→€ 4.05M | 8.1% |
| Layer | Tech. Premium | Expected Loss | Expenses | Allocated SCR | Return on Capital |
|---|---|---|---|---|---|
| Layer 1 above hurdle | € 3.60M | € 1.47M | € 0.52M | € 8.5M | 18.9% |
| Layer 2 | € 4.20M | € 1.40M | € 0.61M | € 12.0M | 18.3% |
| Layer 3 | € 2.80M | € 0.80M | € 0.41M | € 11.5M | 13.8% |
| Layer 4 below hurdle | € 3.20M | € 0.80M | € 0.46M | € 17.5M | 11.1% |
Behind these screens, four capabilities — and one principle: everything traceable.
Complete workflow on vendor cat models (AIR, RMS). All reinsurance structures — XL with reinstatements, Quota Share, Stop Loss, Aggregate covers with annual deductibles, hybrid — with marginal impact analysis and roll-back per layer, event, peril and model.
Gross and net-of-retrocession accumulation with marginal contribution analysis. Full roll-back capability per layer, event ID, peril and vendor model. Full OEP and AEP curves, gross and net of retrocession.
Portfolio optimisation under solvency constraints. Share adjustment to optimise capital cost and return on equity — a direct link between pricing decisions and capital impact.
Recipe system with a visual dependency graph for full traceability and instant replay. Change a source, an FX date or a catalogue — recompute in one click — only what changed. Every number auditable.
You never draw this graph — VinkeRE generates it with every analysis.
Every analysis in VinkeRE automatically produces a Recipe — the full record of its sources, adjustments and structure. The graph below is simply that Recipe made visible. And because the Recipe knows everything, repricing is trivial: update an FX date, apply an inflation scenario, or price the 2027 structure on 2026 YELTs — the engine replays the Recipe and recomputes only what changed. That is exactly what the scenario switches in the cockpit above are doing.
Full codebase held in secure escrow — guaranteeing continuity and complete source access under agreed conditions.
Client code review welcome. Model validation supported by full result convergence against the incumbent vendor engine, recipe traceability, and reproducible seeds stored with every YELT.
Licence & maintenance model — the engine evolves alongside your team, not instead of it.
I have spent over a decade in reinsurance — across SCOR, MS Re, Sompo and Toa Re Europe — pricing treaties, building workflows, and seeing first-hand where actuarial tools fall short. At some point, I stopped waiting for better tools and started building them myself.
You read the terrain, respect uncertainty, trust your preparation, and know when to push — and when to turn back. Vinkery was built with that mindset: rigorous, practical, and designed for decisions where the cost of being wrong is real.
Join us in Zurich, working directly with the founder on reinsurance pricing, VinkeRE development, and R&D on AI-driven actuarial workflows. Master's in actuarial science, advanced Python, and the ambition to build — not just analyse.
Whether you're exploring VinkeRE for your team, interested in joining us, or looking for senior actuarial advisory — we'd like to hear from you.