Скачать или смотреть LLMOps on a Shoestring: Smart Scaling with Vendors, APIs & Microservices

Building scalable LLM applications is complex - but doing it on a startup budget is even tougher. This panel will explore cost-effective strategies for deploying, monitoring, and optimizing LLM-based applications by strategically leveraging vendors, APIs, and architectural components like databases and agents, including:

-The most cost-effective strategies for deploying and scaling LLM-based applications.
-Mission-critical expenses in the early stages and how budgets should evolve as the company grows.
-Evaluating vendors for LLM APIs, databases, and agents while minimizing lock-in risks.
-Leveraging architectural components like RAG, agents, and vector databases to enhance performance.
-When to use third-party APIs versus building in-house LLM infrastructure.
-Best practices for monitoring, observability, and troubleshooting LLM performance in production.
-Optimizing inference costs while maintaining responsiveness and reliability.

You'll leave with actionable insights on how to build a high-performing AI infrastructure without overspending—making smart trade-offs between performance and cost.

This session is part of Startup Boston Week 2025 and is perfect for founders, technical and founding engineers, MLOps professionals, AI product managers, and finance leads looking to build scalable AI products on a budget.

Meet the speakers:
-Nnenna Ndukwe - Developer Relations Lead & AI GTM Advisor, Qodo
-Sri Krishnamurthy - CEO, Quantuniversity
-Lavnish Lalchandani - Senior Engineering Manager
-Kevin Walsh - Co-Founder, Melodi