Senior Risk Data Scientist (Científico de Datos de Riesgo Senior) - Latam

Clara Labs

Full Timesenior

Latin America Posted 7 days ago

Resume Keywords to Include

Make sure these keywords appear in your resume to improve ATS scoring

PythonSQLBootstrapAWSPandasNumPyscikit-learnB2B

Job Description

<div class="content-intro"><h2>Ready to accelerate your career?</h2> <div> Clara is the fastest-growing company in Latin America. We've built the leading solution for companies to make and manage all their payments. We already help over 20,000 large and growing businesses operate with agility and financial clarity through locally issued corporate cards, bill pay, financing, and a powerful B2B platform built for scale. Clara is backed by some of the most successful investors in the world, including top regional VCs like monashees, Kaszek, and Canary, and leading global funds like Notable Capital, Coatue, DST Global Partners, ICONIQ Growth, General Catalyst, Citi Ventures, SV Angel, Citius, Endeavor Catalyst, and Goldman Sachs - in addition to dozens of angel investors and local family offices.  We’re building the financial infrastructure that powers high-performing organizations across the region. We invite you to join us if you want to be part of a fast-paced environment that will accelerate your career and support you to do some of the best work of your life alongside a passionate and committed team distributed across the Americas. </div></div><h1>Senior Risk Data Scientist — Credit Risk Modeling</h1> <h3>Job Description</h3> <h2>About the Role</h2> We are looking for a Senior Risk Data Scientist to join our Risk Data Science team at Clara. This is a role for someone with deep, hands-on credit risk modeling experience. You will own the full model lifecycle: problem definition, data understanding, feature engineering, model development, validation, deployment, and post-deployment monitoring. You will work directly with Risk to build the models and segmentation frameworks that determine who gets credit, how much, and on what terms. <h2>What You Will Do</h2> Credit Risk Modeling <ul> <li>Design, build, validate, and deploy origination scorecards and behavioral models (line management, collections prioritization, cross-sell) for consumer and SME credit products.</li> <li>Engineer credit risk targets — ever30, ever90, roll rates, DPD migration buckets — and manage observation window design, performance window alignment, and vintage construction end-to-end.</li> <li>Implement and document reject inference methodologies (augmentation, parceling, fuzzy assignment), including sensitivity analysis with and without RI, and obtain Risk sign-off on methodology before development begins.</li> <li>Apply temporal cross-validation as the primary development metric when working with small development samples — understanding why a single train/test split with fewer than 150 bads in the test set produces statistically invalid KS comparisons and knowing how to handle it.</li> <li>Conduct challenger vs. champion evaluations including swap-in / swap-out analysis, bad rate comparison on newly approved accounts, displacement profile analysis, and full economic impact estimation.</li> </ul> Population Segmentation & Portfolio Diagnostics <ul> <li>Build and maintain population segmentation frameworks — bureau hit/no-hit, to enable granular model performance analysis and decision strategy refinement.</li> <li>Produce segment-level performance reports covering KS, AUC, Average Precision, bad rate, and rank order monotonicity across train/test splits, time periods, and population cuts — identifying where models underperform and why.</li> <li>Identify thin-file and underbanked segments with high bureau no-hit concentration and assess their statistical suitability for reject inference inclusion, documenting assumptions and limitations.</li> <li>Monitor PSI at score and feature level, applying bootstrap confidence intervals for small portfolios and decomposing exogenous population drift (macro, policy changes) from genuine model degradation.</li> </ul> <ul> <li> </li> </ul> Platform & Engineering <ul> <li>Develop and maintain model pipelines on Databricks, managing feature engineering, training, experiment tracking with MLflow, and deployment workflows at scale.</li> <li>Write production-grade Python and SQL — reusable libraries, not one-off notebooks; code that data engineers can maintain and Risk can audit.</li> <li>Collaborate with data engineering on feature stores, ETL pipelines, and monitoring infrastructure on AWS/Datbricks.</li> </ul>   Collaboration & Leadership <ul> <li>Work end-to-end with cross-functional stakeholders from problem definition through deployment and post-deployment monitoring, maintaining consistent follow-up and clear communication at every stage.</li> <li>Mentor junior and mid-level data scientists — conducting code reviews, guiding modeling decisions, and helping them develop the judgment to distinguish statistical noise from genuine model signal.</li> <li>Leverage AI tools as part of your daily workflow to accelerate development, automate repetitive analysis, and improve team throughput.</li> <li>Communicate model results, segment diagnostics, limitations, and recommendations clearly to both technical audiences and senior business stakeholders.</li> </ul> <h2>What We Are Looking For</h2> Experience <ul> <li>7+ years of experience in data science, analytics, or a related quantitative field.</li> <li>5+ years of hands-on credit risk modeling experience — origination scoring, behavioral models, or both — with models that reached production and were monitored over time.</li> <li>Experience with population segmentation applied to credit risk: cohort construction, performance by segment.</li> <li>Hands-on experience with reject inference methodology and a clear understanding of its implications for model bias, performance estimation, and regulatory documentation..</li> </ul> Technical Skills <ul> <li>Python — pandas, NumPy, scikit-learn, XGBoost / LightGBM, SHAP; clean, maintainable, production-quality code.</li> <li>SQL — complex queries, window functions, large-scale data manipulation; comfortable working directly with raw transactional data.</li> <li>Databricks and/or AWS — hands-on experience running model development, experiment tracking, and deployment workflows on cloud platforms.</li> <li>MLflow or equivalent — experiment tracking, model registry, reproducible pipelines.</li> <li>Deep understanding of model validation metrics: KS (including confidence intervals), AUC, PSI, Gini, calibration error, rank order monotonicity.</li> <li>Familiarity with data visualization tools (Metabase or equivalent) for communicating portfolio diagnostics to non-technical stakeholders.</li> </ul> Soft Skills <ul> <li>Data storytelling — able to translate segment-level diagnostics, KS confidence intervals, and PSI decompositions into clear business recommendations that non-technical stakeholders can act on.</li> <li>End-to-end ownership — you define the problem, build the solution, handle the edge cases, deploy it, and own the monitoring plan; you do not hand off and move on.</li> <li>Resilience and pragmatism — comfortable operating in ambiguity, prioritizing high-impact work, and making decisions with imperfect data.</li> <li>English fluency — all written communication, documentation, code comments, and stakeholder presentations are conducted in English.</li> </ul> <div class="content-conclusion"><h2 data-start="172" data-end="190">Why join Clara</h2> At Clara, you’ll have the autonomy, speed, and support to make meaningful impact — not just on your team, but on how organizations are run across Latin America. <h4 data-start="354" data-end="368">Who we are</h4> <ul data-start="370" data-end="902"> <li data-start="370" data-end="446"> We’re the leading B2B fintech for spend management in Latin America. </li> <li data-start="447" data-end="568"> Certified as one of the world's fastest-growing companies, a Great Place to Work, and a LinkedIn Top Startup. </li> <li data-start="569" data-end="643"> Passionate about making Latin America more prosperous and competitive. </li> <li data-start="644" data-end="749"> Constantly innovating to build financial infrastructure that enables each of our customers to thrive. </li> <li data-start="750" data-end="837"> Product-led, high-talent-density culture — designed for builders who raise the bar. </li> <li data-start="838" data-end="902"> Proud of our open, inclusive, and values-driven environment. </li> </ul> <h4 data-start="904" data-end="926">What we believe in</h4> <ul data-start="928" data-end="1366"> <li data-start="928" data-end="995"> #Clarity. We say things clearly, directly, and proactively. </li> <li data-start="996" data-end="1065"> #Simplicity. We reduce noise to focus on what really matters. </li> <li data-start="1066" data-end="1135"> #Ownership. We take responsibility and never wait to be told. </li> <li data-start="1136" data-end="1201"> #Pride. We build products and experiences we’re proud of. </li> <li data-start="1202" data-end="1287"> #Always Be Changing (ABC). We grow through feedback, risk-taking, and action. </li> <li data-start="1288" data-end="1366"> #Inclusivity. Every voice counts. Everyone contributes to our mission. </li> </ul> <h4 data-start="1368" data-end="1385">What we offer</h4> <ul data-start="1387" data-end="1837"> <li data-start="1387" data-end="1451"> Competitive salary and stock options (ESOP) from day one </li> <li data-start="1452" data-end="1559"> Multicultural team with daily exposure to Portuguese, Spanish, and English (our corporate language) </li> <li data-start="1560" data-end="1629"> Annual learning budget and internal accelerated development paths </li> <li data-start="1630" data-end="1716"> High-ownership environment: we move fast, learn fast, and raise the bar — together </li> <li data-start="1717" data-end="1770"> Smart, ambitious teammates — low ego, high impact </li> <li data-start="1771" data-end="1837"> Flexible vacation and hybrid work model focused on results </li> </ul> If you’re ready for growth, ownership, and impact — apply now and help us redefine B2B finance in Latin America.   <hr> <h4 data-start="134" data-end="159">Clara’s Hybrid Policy</h4> Claridians in a hybrid mode split their time between working from the office, talking to or visiting customers, or working from home. This hits a balance between bringing people together for in-person collaboration and learning from each other, while supporting flexibility about how to do this in a way that makes sense for each individual and team. We don't enforce a minimum number of days for most roles, but you're expected to spend time at the office organically, and be at the office most days during your ramp-up or when required by your leader.</div>