Mariana F.

Data Scientist

Mariana is proficient in Python and R and has expertise in a range of technologies, including SQL, AWS (S3, SageMaker, Redshift), Git, PySpark, Flask, and PyTorch.

She also has hands-on experience with unit and integration testing, feature store management, and tools like Databricks, Looker, and Metabase. Her well-rounded technical skill set allows her to build reliable and scalable data solutions.

Mariana has a strong background in finance and marketing and has transitioned into the tech industry, where she has four years of experience. Her expertise spans analytics, credit modeling, and recommendation systems, driven by her passion for data and statistics. She excels in creating well-crafted, data-driven products that deliver meaningful insights and results.

With a unique combination of analytical acumen and technical proficiency, Mariana is adept at driving data-driven decisions and building impactful products, making her an asset to any data-centric team.

Main expertise
  • Apache Spark
    Apache Spark 5 years
  • AWS
    AWS 5 years
  • Data Science 5 years
Other skills
    Mariana
    Mariana F.

    Brazil

    Get started

    Selected experience

    Employment

    • Data Scientist

      QuintoAndar - 7 months

      • Increased 10% email click-through rate by training, optimizing, and deploying a deep learning two-tower content-based recommendation model.
      • Trained and optimized a deep learning collaborative filtering recommendation model for email campaigns.
      • Responsible for planning 2023 bets, interviewing and hiring, onboarding three new hires, and documenting legacy decisions from previous owners of recommendation models at the company.
    • Data Scientist

      QuintoAndar - 1 year 8 months

      • Improved credit approvals from 20% to 50% by developing and deploying a gradient-boosting classification model.
      • Increased calculation accuracy by 30% by automating the calculation of Net Present Value (NPV) from Excel to Python.
      • Decreased manual analysis from 20% to 10% in offline tests by creating a fallback logistic regression model for credit approvals.
    • Data Analyst

      QuintoAndar - 2 years

      • Improved accuracy in the estimation of supply availability duration by 20% by developing a Survival Analysis method, impacting Unit Economics estimation.
      • Zeroed differences in numbers of the main business metrics across the company by developing a standardized metrics calculation in Metrics Governance initiatives.
      • Freed 25% of work time from one analyst by translating the calculation of business targets from Excel to SQL.
      • Provided reports of main business metrics for Series D and Series E investment rounds.
    • Data Analyst

      CUBOCC - 2 years

      • Developed a report template and process for analysis of marketing strategies of competitors across the company.
      • Developed marketing campaign results template standardized cross-company to automate performance analysis.
    • Controller Assistant

      Continental - 3 years

      • Calculated yearly and monthly forecast estimation for financial statements within 15% of deviation from actual values.
      • Developed dashboard templates and trained teams across countries for reporting.

    Education

    • Standalone courseData Scientist

      Udacity · 2019 - 2019

    • Standalone courseData Science

      Tera · 2019 - 2019

    • BSc.Statistics

      University of São Paulo · 2018 - 2021

    Find your next developer within days, not months

    In a short 25-minute call, we would like to:

    • Understand your development needs
    • Explain our process to match you with qualified, vetted developers from our network
    • You are presented the right candidates 2 days in average after we talk

    Not sure where to start? Let’s have a chat

    First developer starts within days. No aggressive sales pitch.