← back
AI System Design: From Idea to Production - Apoorva Joshi, MongoDB
Takeaway
Derive AI architecture and evaluation criteria from explicit user needs, constraints, and data freshness requirements.
Summary
- Joshi proposes four phases for AI system design: product requirements, system architecture and data, evaluation and monitoring, and optimization for cost, latency, and reliability.
- A hypothetical insurance-claims assistant starts from reviewer delays and measurable outcomes rather than prescribing an agent architecture upfront.
- Cloud and vendor restrictions, required physician involvement, human review of denials, and performance budgets determine the system's permitted autonomy.
- The data strategy maps clinical guidelines, coverage policies, and claims history to storage formats and refresh cadences; an example goal reduces urgent review time to one hour within 90 days of launch.
system-designrequirementsproduction-ai
Original description
Writing code is no longer the hard part. AI can do that. In a world where AI writes the code, the most valuable skill an engineer can have is knowing what to build. Most AI systems never make it to production because of bad decisions made earlier in the process. The pressure to ship fast, the hype around AI, and the lack of a structured approach all push engineers toward building before they've thought through what they're building and why. In this talk, you'll learn a structured framework for making the decisions that get AI systems to production. You'll learn how to identify the business problem, define success metrics, select the right architecture, define guardrails and evaluation metrics, and know what to optimize for when you ship. We'll apply the framework to a real-world problem as we go, so you can see how it works on an actual AI application. Speakers: Apoorva Joshi (MongoDB): Apoorva is currently a Staff AI Developer Advocate at MongoDB. She has a diverse engineering background with a Bachelor’s in Electrical Engineering, a Master’s in Computer Engineering, and several years of experience as a data scientist, applying AI to problems in the cybersecurity space. She now uses that applied AI expertise to help data science and engineering teams at large enterprises and startups build production-grade AI applications with MongoDB and Voyage AI. LinkedIn: / apoorvajoshi95 GitHub: https://github.com/ajosh0504