TechBuzz Meetup: Building Reliable AI Systems
Sigma Software, UzPythonJobs and UzDjangoDevelopersCommunity invite you to TechBuzz Meetup: Building Reliable AI Systems — our first online technical meetup for developers interested in AI engineering.
This meetup will focus on the practical side of building AI systems that not only look impressive but actually work reliably in production: from data pipelines that power agents to monitoring, evaluation, and debugging LLM features.
Details
📅 Date: July 18
🕓 Time: 16:00 (Tashkent time)
📍 Location: Online via Microsoft Teams. After registering, you will receive a link to the meetup.
🗣 Language: Uzbek
🎟 Register to secure your spot!
Speakers & Topics
Speaker #1
Nozir Tukhtamurodov — Senior Software Developer at Sigma Software
Nozir is a Software Engineer with around 7 years of experience across EdTech, FinTech, and AI. His professional interests include high-load applications, multi-agent systems, and building scalable software solutions for complex domains.
Topic: Your Agent Is Only as Good as Your Pipeline — Why Data Comes First
In this talk, Nozir will explain why good data pipelines often decide whether your AI agent works or fails.
You will learn about:
• The math of failure: how small data errors grow into big problems in multi-step agent workflows
• Why AI agents need better data than dashboards and reports ever did
• 3 common failure stories: duplicate records, outdated data behind confident decisions, and bugs that cannot be traced back to the source
• A simple rule: an agent must earn its autonomy — with a practical checklist to use before giving an agent real power
Speaker #2
Bekzod Khudayberdiyev — Software Engineer at Accenture
Bekzod specializes in building backend services, microservices, and RAG/agentic systems. He has a strong interest in delivering reliable, real-world AI solutions that can be monitored, evaluated, and improved in production.
Topic: LLM Observability & Evaluation in Practice
In this talk, Bekzod will cover the practical side of monitoring and evaluating LLM-powered systems.
You will learn about:
• Tracing and logging LLM calls: what to capture and why, including latency, tokens, cost, and prompts
• Evaluation approaches: golden datasets, LLM-as-a-judge, and human feedback loops
• A production chatbot case study, including catching a real regression
Join us to learn, ask questions, and connect with developers in Uzbekistan.
About the organizers
Sigma Software is a global IT consultancy that has been providing top-notch IT services and turnkey software solutions for businesses of all sizes, from start-ups to enterprises, for over 20 years. Over 300 customers, including Fortune 500 companies and technology start-ups, rely on Sigma Software to support their digital transformation. With a presence in over 20 countries, Sigma Software brings together a diverse range of engineering expertise and a shared passion for developing world-class software.
UzPythonJobs is a Telegram channel that helps developers in Uzbekistan to stay informed about job opportunities, industry news, and local tech initiatives.
UzDjangoDevelopersCommunity is a community for Uzbek Django developers to share knowledge, ask questions, exchange experience, and support each other in growing as software engineers.
Together, we aim to create more spaces for practical knowledge sharing, professional growth, and stronger connections within the developer community in Uzbekistan.