6 days ago
Warsaw, PolandSenior
Responsibilities
- Build and ship credit-scoring and merchant-underwriting models end-to-end, including feature engineering, training, validation, deployment, and post-production monitoring.
- Engineer point-in-time-correct features for BNPL and MCA decisioning and partner with Data Engineering on source contracts and freshness SLAs.
- Design and operate agent loops and orchestration graphs with context, tools, evaluations, guardrails, observability, and cost controls.
- Implement model versioning, model CI/CD, champion/challenger deployment, shadow deployment, and serving workflows.
- Monitor production models for drift, stability, and performance degradation, then trigger retraining, recalibration, or rollback when thresholds are breached.
- Build validation and backtesting harnesses within the MLOps pipeline and produce audit-ready evidence packs for independent review.
- Design explainable and regulatory-defensible models and document features, decisions, and validation evidence for CCD2 assessment and audits.
Requirements
- Hands-on experience in applied machine learning at scale across feature engineering, training, validation, deployment, and monitoring.
- Experience with model validation methods including backtesting, population stability index, Gini/KS discrimination, and calibration.
- Experience with MLOps, model versioning, model CI/CD, champion/challenger or shadow deployment, drift detection, and automated alerting.
- Strong point-in-time-correct feature engineering and feature-store consumption experience.
- Familiarity with explainability techniques such as SHAP and LIME and with documentation for regulatory defensibility.
- Fluency in Python and SQL, with Spark and Databricks experience for large-scale model training and serving.
- Experience designing agentic tooling, loops, and graphs, including context and tool design, guardrails, and auditable evaluation outputs.
- Solid working knowledge of LLMs and fluent English at B2+ level.
- Preferred experience includes independent model validation, CCD2 or consumer-credit regulation, MCA underwriting, merchant-risk data, fraud detection, IFRS 9 expected-credit-loss modeling, or real-time low-latency model serving.
Benefits
- Hybrid work, ideally one day per week in the Warsaw or Kraków office.
- B2B contract.
- Opportunity to work on a greenfield, high-stakes fintech platform with strategic impact on AI-native ML at InPost.
- Modern technology stack, experimentation with new AI tooling, and a collaborative environment emphasizing autonomy and accountability.
Tech Stack
Categories
About InPost
InPost is a public logistics company listed on Euronext Amsterdam that builds and operates an out-of-home parcel delivery network for e-commerce. Its network includes more than 50,000 out-of-home pick-up points, including over 27,000 parcel lockers, plus to-door courier and fulfillment services for online merchants across Poland, the UK, Italy, Iberia, France and Benelux. Headquartered in Luxembourg, it acquired French operator Mondial Relay in 2021.