esolutions.tech

We’re happy to see our colleague Roberto Comsa speaking at the upcoming Big Data Meetup in Bucharest.
At eSolutions.tech, we always appreciate opportunities to share knowledge, exchange ideas with others in the industry, and be part of the local tech community. It’s great to see people openly talking about real challenges, lessons learned, and the work happening behind the scenes.
If you’re into data and engineering, it’s definitely worth joining the conversation and meeting others in the field.

Why does data validation matter so much in CDC streaming pipelines? Roberto Comsa, Data Engineer at eSolutions.tech, will explore this scenario through real production systems.

Working with NiFi, Kafka, Spark, Delta Lake, and MinIO, he’ll walk through how CDC data flows from ingestion to consumption and how small issues in streaming pipelines can quietly impact data quality over time. The focus will be on practical validation strategies, early detection of drift, and building trust in continuously moving data.

📍 June 9 | ING Hubs Romania

🎟️ Free entry

👉 Looking to join? More details and RSVP here: https://www.meetup.com/bucharest-big-data-meetup/events/314642829

In the rush to move to agentic workflows, companies may fall into a familiar trap: embedding system-specific logic directly into the application layer. The result? A spaghetti integration where changing a model or updating a database schema could break the entire pipeline.

The Model Context Protocol (MCP) offers a more resilient architectural path. By decoupling reasoning (the model) from ex*****on (the connectors), MCP allows for a modular AI stack:

🔹 Model agnosticism: possibility to swap LLMs without rewriting the integration logic.
🔹 Separation of concerns: the model focuses on intent, while MCP connectors handle the technical "how-to" of accessing APIs or internal databases.
🔹 Scalability: adding proprietary connectors as the environment evolves, without adding complexity to the core agent logic.

We dive deeper into why MCP is becoming the standard for enterprise-grade AI architecture here: https://www.esolutions.tech/mcp-enterprise-ai-integration