Machine Learning Engineer
Lucas is a skilled Data/ML Engineer with over five years of experience in data engineering, data science, machine learning, and Python.
He has a proven track record of designing and implementing various ML and data pipelines, including building a data platform from scratch. With experience working in healthcare, real estate, education, and marine engineering, Lucas is a curious and excited generalist who is always eager to learn new things.
He is keen to work for companies where data or artificial intelligence plays a crucial role.
Currently, Lucas is part of an early-stage health-tech company focused on addressing overweight and obesity issues in Latin America.
- Data Engineering 2 years
- Data Science 4 years
- Docker 3 years
- PostgreSQL 2 years
- Scrum 2 years
- Apache Airflow 1 years
BrazilBook Lucas M.
Senior Data Engineer
Liti - 1 year 9 months
- Part of the founding team built the company's highly scalable, robust data platform using Airflow, Fivetran, DBT and BigQuery, cloud-based in GCP;
- Using dbt, built fully traceable, transparent transformation layers (raw/stg/int/mart) that turned data from Hubspot, Calendly, Typeform, and MongoDB into intuitive, insightful data models;
- Using Python and Airflow, developed and orchestrated data pipelines for core business data sources, such as Calendly and Google Sheets;
- Built an EHR system that uses Python, Kubernetes, and Cloud SQL and integrates with BigQuery;
- Reduced health professionals' time spent looking for data in 30 minutes per appointment.
- Apache Airflow
- Data Engineering
- Google Cloud
Loft - Top Brazilian Unicorn startup - Lunicorn - 8 months
- Achieved a series D valuation of 3 billion USD.
- Developed several ETL processes in the company's large-scale data warehouses using Python, Pandas, Databricks, and Spark.
- Conducted A/B tests using 100GBs of navigation data to determine the impact of recommendation models on company sales.
- Optimized a Python Machine Learning recommender system for real estate property visits using bounding box policies, resulting in a 15% uplift in visit schedules.
- Built a multiple regression system using Python, Pandas, Streamlit, and Plotly to allocate real estate brokers across three of Brazil's largest cities, minimizing visit overbookings by over 30%.
- Constructed a sales forecast using a SARIMA model.
- Data Science
- Machine learning
Data Scientist and Instructor
Ada Tech - 1 year
- Created an entire BI platform for the company using Python, Pandas and Streamlit;
- Developed lots of ETL processes, integrating with the company's data warehouse;
- Led two end-to-end data projects;
- Built data pipelines, data models, live analytics dashboards, machine learning models, CI/CD processes and A/B tests, all AWS native, using Lambda, S3, and Code (Commit, Build, Pipeline);
- Worked on machine learning-based sales, scoring the system that leverages data from sources such as WhatsApp and Google Analytics;
- Developed a recommender system used to find the questions that best fit the studying pattern of a given student, using Reinforcement and Unsupervised Learning.
- AWS Lambda
- Machine learning
Reinforcement Learning Researcher
University of São Paulo - 1 year 1 month
- Co-author of AI Research for large ships in the real world, restricted environmental and maneuvring conditions;
- Designed experiments and modeled data representations for training state-of-the-art Asynchronous Deep Reinforcement Learning algorithms using Python, Tensorflow and Redis;
- Developed a robust RL training infrastructure based on Docker that provides scalability and reliability for the heavy Machine Learning operations on the local cluster;
- Created visualization and testing tools for the terabyte, week-long AI training sessions, which allowed significant research advancements.
- Machine learning
Escola Politécnica da USP · 2017 - 2021
Find your next developer within days, not months
We can help you deliver your product faster with an experienced remote developer. All from €31,90/hour. Only pay if you’re happy with your first week.
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
- Share next steps to finding the right match, often within less than a week
First developer starts within days. No aggressive sales pitch.