ML & Data Meetup RWB on December 5: What to Expect and Details

On December 5, the ML & Data Meetup RWB will take place, bringing together top experts from RWB, Avito, and VK Tech to discuss real-world cases of implementing machine learning and artificial intelligence in production. The event will be held in a hybrid format with an offline venue in Moscow and a live online broadcast.

• Date and time: December 5, 13:00.
• Participating companies: RWB, Avito, VK Tech.
• Format: Offline in Moscow + online broadcast.
• Main topics: ML architecture, inference, data pipelines.

Model Architecture and Inference Scaling

How do you bring ML and Data solutions to production when balancing computational constraints, data quality, latency, and strict business requirements? At the meetup, experts will break down the journey from early experiments to pre-training a model with 100 billion parameters, as well as building an inference platform on KServe.

Speakers will demonstrate how to manage model lifecycles, ensure fast scoring and weight updates in real-time ML bidding, and apply mathematical optimization to reduce search spaces in routing tasks. Engineering trade-offs when scaling large ML systems will also be thoroughly examined.

Data Pipelines and Reliable Infrastructure

The event will place a special focus on data engineering: what happens to a data pipeline when event volumes grow, delays occur, or consistency is lost? Through real-world examples, engineers will analyze synthetic data generation for OCR, embedding selection, and building a stateful segmentator using Kafka, Flink, RocksDB, and ClickHouse.

Experts will also cover state troubleshooting, replay mechanisms, deduplication, and observability. A separate session will explore how to seamlessly integrate AI models into product services and build infrastructure designed to handle high-load environments.

Community Value and Final Thoughts

Such industry meetups provide immense value for data scientists and ML engineers, offering direct insights from top-tier tech companies. The event will conclude with a festive atmosphere, networking opportunities, and discussions among professionals who build and maintain production-grade ML and Data systems.