Ben & John'sData Science and Analytics Start a project

Connect the data. Move the numbers.

Ben & John's is a data science and analytics studio in Curitiba, Brazil, working with companies anywhere. Everything on the orange line is something we build.

We work on Curitiba time, UTC−3 all year.

What we build

Six kinds of work, listed in the order data usually needs them. Start with whichever one hurts most today.

  1. Data pipelines and automation

    Scheduled jobs that pull data out of your ERP, CRM and spreadsheets, clean it and keep it in one place. The report someone rebuilds by hand every Monday starts building itself.

    Typical first projects

    • A warehouse fed from ERP or SAP exports
    • A weekly report that assembles and sends itself
    • An API link between two systems that never talked

    First delivery

    1 to 3 weeks

  2. Dashboards and reporting

    One set of numbers for sales, finance and operations, refreshed on a schedule and defined the same way for everyone in the meeting.

    Typical first projects

    • Metric definitions agreed with each team
    • Executive and day-to-day views in Power BI
    • Alerts when a number leaves its usual range

    First delivery

    2 to 3 weeks

  3. Forecasting and machine learning

    Models trained on your own history that estimate what comes next and flag what looks wrong. The error is measured and shown next to every forecast.

    Typical first projects

    • Demand and revenue forecasts
    • Customer segments and churn risk
    • Anomaly detection in transactions

    First delivery

    3 to 5 weeks

  4. Optimization models

    For decisions with more moving parts than a spreadsheet can hold: how much to stock, what to produce first, which route, which price.

    Typical first projects

    • Inventory and replenishment levels
    • Production and staff schedules
    • Pricing and promotion scenarios

    First delivery

    3 to 6 weeks

  5. AI agents and assistants

    Language-model agents that answer questions from your data, read documents and email, and take routine steps inside your tools. A person approves anything that matters.

    Typical first projects

    • An ask-your-data assistant for one team
    • Document and inbox triage
    • Agents that file, update and follow up

    First delivery

    3 to 6 weeks

  6. Websites and web apps

    The surface for everything above: internal tools, client portals and company sites that put the numbers in front of the people who use them.

    Typical first projects

    • Internal tools and portals
    • Interfaces for agents and models
    • Company and product sites

    First delivery

    2 to 4 weeks

A first project, week by week

A three-week pilot is the size we suggest for a first job. Bigger scopes stretch the orange bar and nothing else.

  1. Intro call

    30 minutes, free. You describe the problem and we ask about the data behind it.

  2. Written proposal

    Scope, timeline and one fixed price, in writing, before any work starts.

  3. Pilot build

    Three weeks of building, on your real data.

  4. Handover

    Documentation and a training session. The code, the models and the data stay with you.

  5. Support, if you want it

    Optional. A monthly plan that keeps everything running and adds new analyses as you need them.

Your clock and ours

We're available from 07:00 to 23:00 Curitiba time, weekends included. Curitiba stays on UTC−3 all year, so pick a city to see how much of your working day that covers.

London is 3 hours ahead of Curitiba. We share 8 hours of the working day there, from 10:00 to 18:00 London time.

The top bar assumes a 09:00 to 18:00 working day on your side. Ours is 07:00 to 23:00 in Curitiba, Saturdays and Sundays too.

Meet the partners

Two mechanical engineers from UFPR, both working in data at a global energy company. The name is ours too: Ben is Wanderlei Benhur and John is João. You always talk to the people doing the work.

João Henrique de Cerjat Duarte

João Henrique de Cerjat Duarte

Co-founder, data science and planning

Mechanical engineer from UFPR, the Federal University of Paraná. He has spent three years at a global energy company, first in cost accounting and billing analysis, now in supply planning, where he schedules product replenishment across a worldwide network. His daily tools are SAP, SQL, Snowflake, Power BI and Python.

Wanderlei Benhur Marques dos Santos

Wanderlei Benhur Marques dos Santos

Co-founder, data engineering and operations

Mechanical engineer from UFPR, with a postgraduate degree in data science. At the same global energy company he coordinates four international warehouses, after starting there in data analytics and data engineering, building dashboards and automations and supporting the systems its refineries use worldwide. His daily tools are Python, SQL, Snowflake, Power BI and Tableau.

Based in
Curitiba, Brazil
Time zone
UTC−3, all year
Hours
07:00 to 23:00, seven days a week
Languages
English and Portuguese
Works with
Companies anywhere, remotely
Registered
Brazil, CNPJ 68.942.043/0001-20
Toolbox
Python, SQL, Snowflake, Power BI, Tableau, SAP, LLM APIs, optimization solvers

Got a number that needs moving?

Tell us about it in a few lines. The first call takes 30 minutes and costs nothing.

Hours
07:00 to 23:00 Curitiba time, weekends included.
Where
Curitiba, Brazil. We reply within one business day.

We use these details only to reply to you. Nothing leaves your browser until you send it by email or WhatsApp. Read the privacy policy