Design and implement AI features in strategic products.
Build LLM-based solutions for use cases such as customer support, triage, recommendation, information generation/extraction, and flow configuration/automation.
Develop and orchestrate multi-agent architectures for complex tasks.
Contribute hands-on to experimentation, architecture decisions, and the design of ML and AI solutions for the Expenses business unit products.
Create AI services and production pipelines (APIs, observability, logging, performance monitoring).
Define and maintain quality metrics, using experimentation (A/B or controlled tests) and continuous improvement.
Help ensure security, privacy, and compliance in the use of sensitive data, including best practices for anonymization and governance.
Document technical decisions, trade-offs, and patterns to help scale AI adoption across the company.

Degree in Computer Science, Statistics, Mathematics, Engineering, Physics, or related fields.
Solid experience as a Data Scientist / AI Specialist (minimum 5 years) with production deliveries.
Hands-on experience applying LLMs to product use cases.
Proficiency with data tools and platforms (e.g., Databricks, Python, SQL, Metabase).
Experience building LLM solutions (RAG, fine-tuning) and developing multi-agent architectures with quality measured by Evals and clear metrics.
Experience orchestrating agentic workflows with LLMs using frameworks like LangGraph.
Solid foundation in applied Machine Learning (classical models, metrics, validation, experimentation, and monitoring/drift) to complement building LLM solutions in production.

Flash Card (the beloved pink one!) with flexible benefits: meal, grocery, mobility, health, education, culture, and wellness
Health plan
Life insurance
Extended maternity and paternity leave + daycare allowance
Birthday day off 🎂
Hybrid and flexible work model + home office allowance + in-office experiences
Exclusive partner discounts through the Flash app
TotalPass
Pet care plan with Guapeco

Data Science Specialist

Key skills