Proven Results - Model Development

---

These case studies demonstrate how cipris delivers AI-first model development: solving complex quant challenges fast, accurately, and with regulatory integrity(*).

What the client needed
A FinTech firm operating in low-liquidity commodity markets wanted pricing capabilities for vanilla, barrier, binary and FlexStrip options that remain accurate under stressed volatility, deep in/out of the money, and uneven expiries.

What we did
We built an AI-first exotic derivatives pricing framework leveraging QuantLib, including dual-curve yield models (LIBOR / RFR), robust bootstrapping, implied volatility surfaces with ML outlier detection, and arbitrage-free volatility curves for grain, softs and energy. Models were designed, validated and documented for regulatory compliance.

What the client gained

• Reliable pricing across option types in volatile commodity environments
• Confidence in structured product valuation even under extreme market stress
• Production-ready pricing library backed by rigorous testing and error control
• Transparent, AI-powered model performance under diverse expiry and strike combinations

What the client needed
A global investment bank required an equity-derivatives valuation library with integrated testing to streamline model validation and governance certification.

What we did
Implemented valuation-model testing and performance analysis for equity-derivatives models (C++/Python); produced high-quality LaTeX documentation; and coordinated closely with governance/validation teams to expedite certification.

What the client gained

• Production-ready equity-derivatives library with embedded validation
• Faster certification via coordinated governance workflows
• Comprehensive documentation across products and models
• Operational efficiency without compromising rigor

What the client needed
Trading operations needed advanced interest-rate modeling (exotics) with high-fidelity calibration and robust numerical methods.

What we did
Implemented and calibrated a multi-factor BGM model for exotic rate trades with cap/swaption calibration; delivered a custom C++ library (Sobol/Mersenne Twister RNG, Nelder–Mead optimization, Simpson integration, dual-curve/OIS/CVA discounting, Monte Carlo, LU/Cholesky).

What the client gained

• Sophisticated rate-modeling capability for complex exotics
• Higher valuation precision and risk insight across scenarios
• Confidence in numerical stability and model performance
• A maintainable, production-grade codebase for the desk

What the client needed
A major European bank faced delays and inconsistency in small-business loan approvals due to manual risk assessments across regions.

What we did
Developed an economic-capital model with RAROC-based decisioning; calibrated PD/LGD; rolled out across seven countries; conducted six months of testing; trained 50 managers; and embedded governance/validation.

What the client gained

• 400% increase in automated approvals (to ~70% under delegation)
• Decision times cut from days to minutes/events
• Regulatory-compliant credit-risk modeling and documentation
• Stronger internal capability via training and processes

What the client needed
A major European bank’s ALM committee required monthly forecasts and scenario analysis for interest-rate risk in the banking book to guide balance-sheet strategy.

What we did
Built an IRRBB ALM model with EaR/EVE/sensitivity analysis, new-business forecasting, prepayment (CPR) assumptions, rate–balance correlation, and automated scenarios; validated for regulatory alignment and integrated with existing tools.

What the client gained

• Executive Committee-ready monthly scenario reporting for IRRBB decisions
• Improved foresight on exposure and funding strategy
• Embedded tooling—no manual workarounds
• Simplified oversight via built-in governance/documentation

(*) Some results are based on the founder's extensive experience at top-tier financial institutions.