PySpark Training in Belgrade

U ovom instruktur vođenom, uživo treningu u Belgrade, učesnici će naučiti kako da koriste Python i Spark zajedno za analizu big data, radeći praktične vežbe.

Krajevom ovog treninga, učesnici će moći da:

• Nauče kako da koriste Spark uz Python za analizu Big Data.
• Rade vežbe koje imitiraju slučajeve iz stvarnog sveta.
• Koriste različite alate i tehnike za analizu big data koristeći PySpark.

This instructor-led live training (online or onsite) is designed for data engineers, analysts, and professionals who wish to use Databricks and PySpark to build scalable data pipelines and migrate existing SQL workflows.

This course offers a hands-on introduction to developing scalable data processing and Machine Learning workflows using PySpark. Participants will gain insights into how Apache Spark functions within contemporary Big Data ecosystems and learn to process large datasets efficiently by applying distributed computing principles.

This hands-on, three-day program is dedicated to designing and tuning high-efficiency data processing pipelines that leverage PySpark, Pandas, and Polars within Kubernetes-based ecosystems.

Learners will gain a working knowledge of how Spark applications are orchestrated on Kubernetes, with a specific focus on how configuration choices at the application layer directly impact throughput, scalability, resource utilization, and operational expenditure. Key areas of optimization covered include executor dimensioning, memory distribution, dynamic resource allocation, partitioning methodologies, shuffle mechanics, mitigating small-file issues, and streamlined Parquet data handling.

Additionally, the curriculum tackles frequent bottlenecks associated with Pandas, such as memory constraints and out-of-memory exceptions, while presenting Polars as a high-performance alternative for specific processing tasks. Through interactive lab sessions, participants will learn to pinpoint performance and memory inefficiencies, evaluate various configuration approaches, and implement optimization strategies for realistic ETL and machine learning use cases.

The core objective of this training is practical decision-making: mastering the identification of system bottlenecks, selecting the right tool for the job, configuring Spark for maximum efficiency, and striking a balance between processing speed and infrastructure resource costs.