Data Engineers in Toronto

Data Engineers in Toronto August 2026 Semimonthly Meeting

Topic: Running GitHub Actions Offline: Debug CI/CD workflows right from your local machine

Abstract:
Ever wanted to try out your GitHub Actions without waiting for the cloud to spin up the runners? Well, you can! Testing GitHub Actions locally lets you build, test, and debug your CI/CD workflows right from your machine. Through a tool such as act, you can simulate GitHub’s environment for your workflows locally using Docker. This means no more pushes of dozens of commits just to see if your syntax in the YAML file is wrong. You can make changes, execute, and validate your automation as many times as you want, even on a plane where there's no Wi-Fi. Just install Docker, install act, load your secrets locally, and run commands like act push to test your events in the workflows. That's faster, lightweight, and the greatest productivity booster. Taking GitHub Actions offline is not about isolation; it's about freedom and control over your automation pipeline.

Speaker: Steve Yonkeu, Software Engineer & Microsoft MVP

Speaker Profile:
Steve Yonkeu is a backend software engineer dedicated to delivering excellence in architecture, performance, and modularity. For over 5 years he has been leading teams and projects and open source communities worldwide. Steve is also the founder of Django Cameroon and Python Cameroon communities today. He is also an organizer at PyCon Africa and speaker at PyConUS.

Data Engineers in Toronto September 2026 Semimonthly Meeting

Topic: Building Collaborative AI Agents Using AI Foundry Agent Service

Abstract:
As enterprise AI adoption grows, many real-world scenarios can no longer be solved by a single conversational agent. Instead, they require multiple specialized agents that collaborate, share context, and delegate responsibilities intelligently. This session focuses on how to design and implement multi-agent architectures using Azure AI Foundry’s Agent Service.

We’ll start by breaking down when and why multi-agent systems make sense, then walk through a practical scenario where agents are assigned clear roles such as data retrieval, reasoning, and action execution. Using AI Foundry Agent Service, you’ll see how these agents are created, connected, and orchestrated to work together while maintaining context and control.

Rather than focusing on theory, the session emphasizes practical design patterns, architecture decisions, and implementation considerations, including agent boundaries, communication flows, and scaling agent-based solutions safely in enterprise environments.

By the end of the session, attendees will have a solid understanding of how to move from a single-agent approach to a collaborative, multi-agent model and how to apply these patterns using AI Foundry in their own solutions.

Key Takeaways:
- Understand when multi-agent architectures are the right choice
- Learn how to design role-based agents using AI Foundry Agent Service
- Explore real-world coordination and orchestration patterns
- Gain practical guidance for scaling and governing multi-agent systems
- Leave with a reusable architecture blueprint for enterprise AI agents

Speaker: Mitul Tailor, Snr. Data & AI Engineer

Speaker Profile:
am a Senior Data & AI Engineer at Long View Systems, specializing in Azure services, data engineering, and AI-driven solutions. With expertise in Azure Data Factory, Fabric, Copilot Studio, and data migration, I have successfully delivered scalable and efficient data solutions.

I hold a Post Graduate Certificate in Generative AI, a PG Diploma in ML & AI, and a Bachelor's in Computer Science. I am DP-600 & AI-102 certified, demonstrating my expertise in Microsoft Azure Enterprise Data Analyst solutions. Additionally, I have a strong understanding of Retrieval-Augmented Generation (RAG) AI and Azure AI Foundry, exploring how these technologies can be leveraged for intelligent automation and data-driven insights.

Recognized by clients for my clear and precise development work, I am passionate about automation, AI, and robotics. Outside of work, I love to travel and have a deep interest in cars.

Data Engineers in Toronto September 2026 Semimonthly Meeting

Topic: Streaming your database - Easier said than done?

Abstract:
Do you need to stream your database? So did I! Here’s what to expect.

In my journey as a contractor for a major cybersecurity company, I implemented a Change Data Capture (CDC) system that aimed to seamlessly stream client databases into a centralized cloud solution for analytics and enhanced features. However, after deploying the code, I quickly realized that the reality was far more complex than anticipated.

Despite using an off-the-shelf product, I encountered unexpected challenges related to scaling, data formats, integrations, and deployment—issues that took me by surprise and required significant adjustments.

Over the course of 18 months, my team and I navigated these hurdles to create a robust system designed for scalability and resilience against failures.

In this talk, I’ll share key lessons learned and practical tips for developers and architects embarking on similar projects. You'll gain insights into preparing for potential surprises and overcoming challenges in building a reliable CDC system.

Speaker: Sigal Shaharabani, Technical Group Leader & Couchbase Ambassador

Speaker Profile:
Sigal Shaharabani is a Technical Leader and a Group Leader at Tikal and a Couchbase ambassador, with a great passion for backend and data systems. She started her technological career way back in 1996, working for the government, a variety of corporations, and small start-ups. Her favorite programming language is Kotlin, and will love any excuse to talk about it. In her spare time she enjoys swimming and Israeli folk dancing.