Jobs — Novacture

Academic and operational objectives - Design, develop and compare several approaches to:

• Establish a claim incidence law segmented by construction characteristics (residence type, construction duration, sum insured level, etc.)
• Describe and calibrate the decennial emergence pattern of claims (non-uniform emergence profile)
• Model the activation date of contracts and quantify residual delays for overdue policies
• Integrate these estimates into the IFRS 17 process: LRC cash-flow projection (Best Estimate), uncertainty measurement (Risk Adjustment), and contribution to quarterly closings

Methodological approach

• Risk models & ruin theory (Cramér-Lundberg, Sparre Andersen) for expectation and uncertainty measurement
• Collective frequency-severity models (GLM): parsimonious variable selection, specification tests, stability vs. explainability
• Survival methods (activation & overdue): Kaplan-Meier, censored quantile regression (CRQ), AFT (Weibull, log-logistic) and Cox for temporal dynamics and effective exposure
• Machine learning methods (gradient boosting, including quantile regression) for comparison
• 10-year emergence profile: construction of a decennial projection vector of future claim costs with sensitivity analyses (shift/flattening of peak, asymmetry)

IFRS 17 integration - Academic evaluation (calibration, information criteria, cross-validation) and integration in the IFRS 17 chain:

• LRC projection (post-activation cash-flows)
• Risk Adjustment (uncertainty quantification)
• Contribution to quarterly closings: Analysis of Change (AoC), volume updates, experience variances

Internship workflow

• Literature review (risk theory, censored survival, emergence profiles, IFRS 17)
• Mathematical design & model development (formulation, specification, calibration, implementation, validation)
• Comparative analysis of methodologies with justification
• Presentation and validation of results with the pool-leading company
• Participation in one or more IFRS 17 quarterly closings

Desired profile

• Master’s degree in Actuarial Science, Statistics or Data Science
• Good knowledge of probability, risk theory and survival models
• Interest in IFRS 17
• Scientific programming skills (Python). Rigor, analytical mindset, interest in methodological analysis.

Academic purpose - Design and test an ALM model for Universal Life (UL with COI) to:

• Project equities, rates and spreads under coherent best-estimate trajectories in the physical measure: drifts anchored to empirical risk premia and term structure (forward curves), with multi-asset dependencies
• Model the UL mechanics (account value, policy/fund fees, COI, surrenders, deaths) and dynamic asset↔liability interactions (liquidity, rebalancing, transaction costs)
• Estimate the stochastic Best Estimate of cash-flows and analyze sensitivities to market, management and behavioral assumptions
• Improve the VFA algorithm for the entity share in coherence with underlying asset dynamics

Scientific approach (comparative, clear and rigorous)

• Market models under P and term structure: parsimonious diffusions (e.g. GBM/Heston-lite for equities, Vasicek/CIR for rates, spread processes); estimated cross-asset correlations and simple no-arbitrage constraints.
• Portfolio management and rebalancing rules: target allocation, calendar/threshold rebalancing, cash-bucket, transaction costs, liquidity and weighting constraints.
• UL mechanics on the liability side: account value algorithm (interest, fees, COI), partial/full surrenders, death benefits, penalties, fund switches; policy→fund mapping and portfolio aggregation.
• Dynamic policyholder behaviors: P-measure laws for surrenders/contributions/switches depending on performance, drawdown, duration and policyholder characteristics.
• ALM-liquidity coupling: asset→liability→asset chain (outflows, claims, asset sales); liquidation rules, redemption windows, robustness tests.
• VFA: entity share - methodological adjustments based on asset paths and policyholder participation.

Evaluation framework and IFRS 17 integration - Evaluation using academic ALM criteria:

• Empirical calibration and diagnostics (residuals, out-of-sample/time-split stability, stylized-fact preservation)
• Numerical stability and conservation of mass (asset-liability flows)
• Sensitivities & stresses: volatility/spread shocks, liquidity crises, surrender/behavioral shocks, management-rule variations
• Focused IFRS 17 reading: Best Estimate under P, VFA coherence through entity share, and alignment with closing requirements without burdening the ALM focus

What you will do in practice

• Targeted literature review (life ALM, UL, physical measure, dynamic behaviors)
• Parameterization and use of an existing ESG; estimation of risk premia, volatilities and correlations
• Python development of an asset-liability projection engine; management rules and liquidity
• Behavior calibration, Monte Carlo simulations, sensitivities/stresses and structured analyses
• Contribution to refining the VFA entity share in the production engine

Desired profile

• Master’s degree in Actuarial Science / Financial Engineering / Statistics
• Strong foundations in stochastic processes & portfolio management
• Advanced Python practice, interest in IFRS 17 (VFA)

Skills to acquire

• Financial modeling (P-measure, term structure, dependencies)
• Mixed life insurance (UL + COI)
• Dynamic ALM (asset-liability feedback, liquidity, management rules)
• IFRS 17 / VFA (Best Estimate, entity-share logic)
• Python engineering (vectorization, quality, auditability)

Working language: French (academic English appreciated).

Academic purpose - Design and implement a coherent modeling framework enabling:

• Modeling mortality and early-repayment behavior by age, duration and credit type
• Building pricing bases structured by entry age and contract duration, for tariff grids and loss-ratio review
• Computing pure premium and commercial premium (expenses, margin, taxes) for dynamic pricing
• Projecting benefit and premium flows for IFRS 17 reserving (LRC/LIC) and uncertainty measurement (Risk Adjustment)
• Documenting a replicable, auditable, automatable framework in Python

Scientific approach (general and rigorous)

• Structuring and cleaning borrower-insurance data, defining technical assumptions (mortality, behavior, rates)
• Explanatory modeling of mortality and early-repayment intensities based on portfolio characteristics
• Building and validating pricing bases age × duration ensuring technical and economic consistency
• Sensitivity analyses and robustness studies on key assumptions
• Industrialization of calculations in Python (pricing modules, projection, calibration, reporting) with reproducible scripts and built-in controls

IFRS 17 evaluation and integration

• Model validation on historical data: temporal consistency, stability and robustness
• Transferability to IFRS 17 closings: service-flow projection, documented Risk Adjustment, complete traceability of assumptions
• Production of deliverables compatible with actuarial closing and reporting processes

What you will do in practice

• Targeted literature review (credit-life pricing, borrower behavior, IFRS 17 reserving)
• Design of the projection model and pricing modules
• Python development of a prototype actuarial tool (pricing & projection engine, automated pricing bases by age and duration)
• Tests, sensitivity analyses, validation and presentation of results to actuarial and product teams

Desired profile

• Master’s degree in Actuarial Science / Statistics / Data Science
• Interest in life pricing, survival models and IFRS 17 work
• Proficiency in Python (pandas, numpy, scikit-learn) & good Excel practice
• Rigor, analytical mindset, interest in IFRS 17

Technical objectives

• User management: authentication, RBAC (Admin/Analyst/Reader), collaborative workspaces
• Subscription management: rights, usage tracking, plan quotas/limits
• Multi-tenant architecture: data isolation, invitations, permissions, configurations
• Security & compliance: audit logging, rate limiting, 2FA, GDPR (export, right to erasure)
• Application infrastructure: real-time notifications, dashboards, webhooks, admin panels
• Plugin system: extensible architecture for SaaS modules

Tech stack

• Frontend: React/Next.js, TypeScript, Tailwind CSS
• Backend: FastAPI (Python), PostgreSQL, Redis, Firebase
• Infra: Docker, Kubernetes, Scalingo, CI/CD (GitHub Actions, SonarCloud)

Required skills

• Python/FastAPI, REST APIs
• Relational SQL modeling
• Git, software testing, maintainable & documented code

Nice-to-have skills

• DDD, Clean Architecture, CQRS
• Front-end & back-end TDD
• DevOps & Cloud, containerization
• API security (JWT, OAuth, RBAC)
• Multi-tenant SaaS, modular systems

Mission workflow

• Functional analysis & DDD design
• Iterative development, code reviews, pair programming
• CI/CD setup & continuous deployment (Scalingo)
• TDD, refactoring, documentation
• User acceptance testing, production deployment, knowledge transfer

Desired profile

• Final-year engineering school or Master’s in Computer Science
• Strong interest in SaaS and product-oriented development
• Autonomy, rigor, end-to-end ownership

Perspectives

• Distributed computing, ML/AI integration
• Event-driven architecture, cloud-native migration
• New analytical components

Conditions

• Compensation depending on profile
• Partial remote work possible
• Individual technical mentoring
• Possibility of full-time hiring

Application

• CV, GitHub/portfolio, short cover letter
• Interview: show and explain the code of a project