Data Scientist - Predictive Maintenance

Tractian; São Paulo; Presencial; BRL 160.000 - 200.000; Tempo integral

Resumo da oferta

A leading tech company in São Paulo seeks a Data Scientist specializing in Predictive Maintenance. Your role involves developing advanced algorithms and predictive models to enhance industrial operations. You'll analyze sensor data, collaborate with engineering teams, and improve data pipelines. An ideal candidate has a strong background in machine learning and experience with time-series data, as well as proficiency in Python. This role offers a unique opportunity to impact the efficiency of industrial processes.

Qualificações

• Expertise in machine learning, time-series analysis, and anomaly detection.
• Proficiency in Python and common data science/ML libraries.
• Solid understanding of signal processing concepts.

Responsabilidades

• Develop predictive maintenance algorithms using machine learning techniques.
• Analyze sensor data streams to identify failure patterns.
• Collaborate with engineers to improve data pipelines.

Conhecimentos

Machine learning

Time-series analysis

Anomaly detection

Python

Data visualization

Signal processing

English (advanced)

Ferramentas

NumPy

pandas

scikit-learn

PyTorch

Descrição da oferta de emprego

Data Science at TRACTIAN

The Data Science team at TRACTIAN focuses on extracting valuable insights from vast amounts of industrial data. Using advanced statistical methods, algorithms, and data visualization techniques, this team transforms raw data into actionable intelligence that drives decision-making across engineering, product development, and operational strategies. The team constantly works on optimizing prediction models, identifying trends, and providing data-driven solutions that directly enhance the company’s operational efficiency and the quality of its products.

What you'll do

As a Data Scientist - Predictive Maintenance at TRACTIAN, you will work at the intersection of advanced data science and industrial operations. Your mission is to develop cutting-edge algorithms and predictive models to monitor and predict equipment failures before they occur, optimizing asset reliability and reducing downtime. You’ll face complex challenges involving large-scale time-series data, real-time data processing, and machine learning applications, while collaborating closely with engineers and laboratory teams to ensure our predictive maintenance solutions remain industry-leading.

Responsibilities

• Develop predictive maintenance algorithms using machine learning techniques for time-series data.
• Analyze sensor data streams to identify patterns that predict equipment failure.
• Research and stay up to date with academic literature and state-of-the-art condition monitoring techniques, translating relevant advances into practical solutions.
• Collaborate with engineers to improve data pipelines and enhance model accuracy.
• Build scalable, real-time models for low-latency predictions.
• Create diagnostic tools that enable data-driven maintenance decisions.
• Work with the laboratory team to design experiments and develop failure datasets using real machinery to validate hypotheses, develop new models, and optimize existing ones.
• Continuously refine models based on real-world performance, experimental results, and feedback.

Requirements

• Expertise in machine learning, time-series analysis, and anomaly detection.
• Proficiency in Python and common data science and ML libraries (e.g., NumPy, pandas, scikit-learn, PyTorch).
• Solid understanding of signal processing concepts and hands‑on experience with industrial sensor data (e.g., vibration, current, temperature, pressure).
• Ability to read, interpret, and apply insights from academic literature and state-of-the-art research in condition monitoring and fault diagnosis.
• Experience designing experiments to validate hypotheses and benchmark models.
• Strong problem‑solving skills and ability to handle noisy, high-dimensional data.
• Advanced English.

Bonus Points

• Familiarity with both academic research and real-world applications in condition monitoring, fault diagnosis, and prognostics (e.g., vibration-based methods, model-based vs. data-driven approaches).
• Experience translating academic methods into robust, production-ready algorithms.
• Prior experience working in industrial or manufacturing environments.