TechPrufer
Get started
AI Engineer
AI Engineer logoAI Engineer · Feature trail

Distributed Systems

5 videos · Sep 2026 – Oct 2026

What this trail covers

This post contributes to the understanding of agent-to-agent communication and collaboration by highlighting the limitations of current approaches and proposing a new framework for distributed AI systems. It emphasizes the need for real-time ordered transport, persistence, runtime binding, agent-first abstractions, and enterprise governance to enable true multi-agent collaboration. The proposed solution, Band, aims to provide a global interaction layer for agents, moving beyond simple messaging.

All AI Engineer trails
  1. Beyond RAG: Build a Relational Context Engine from Scratch — Peter Werry, UnblockedThis talk introduces a 'Relational Context Engine' that goes beyond traditional RAG by transforming natural language queries into structured database queries. It addresses the limitations of similarity search for engineering context (PRs, tickets, chat) which has inherent structure (authors, dates, relationships). The engine is built in four parts: schema discovery (sampling documents to infer types and enums), identity resolution (mapping users to their GitHub logins), query planning (forcing tool calls and including dates), and validation (using allow lists for operators and sanitizing metadata for tenant isolation). A retry loop handles execution errors, and the system demonstrates the need for full-text search when exact matches fail. The same engine powers a chat agent, and Q&A covers scaling schemas, continuous ingestion, and hybrid search.
  2. Why Your AI Agents Can't Talk to Each Other (Yet) — Vlad Luzin, BANDThis talk argues that current approaches to AI agent communication are insufficient, often relying on manual routing or basic scripting. It identifies the single-agent bottleneck (confirmation bias, attention dilution) and explains why existing messaging protocols (Slack, MCP, A2A) fail to enable true agent-to-agent collaboration. The core technical contribution is framing agent connection as a distributed systems problem, requiring real-time ordered transport, persistence, runtime binding, identity, and observability. The talk introduces BAND as an interaction layer designed to address these challenges, demonstrating agents onboarding and collaborating across users.
  3. Move Fast and Don't Break Things: Scaling Databases for the AI Era — PlanetScaleThis talk extends the understanding of distributed systems and database scaling for AI workloads by detailing how PlanetScale's architecture addresses the challenges posed by AI agents. It elaborates on principles of isolation, redundancy, sharding, back pressure, and decoupling. Crucially, it introduces novel agentic capabilities for infrastructure management: agents designing Vitess sharding schemes from VSchema JSON, Traffic Control for resource budgeting to gracefully handle traffic surges, and database branching with deploy requests and one-click reverts for safe schema changes. The core thesis is that good developer experience primitives directly translate to effective AI agent infrastructure.
  4. The Human Is an Async API — Melanie Warrick, TemporalThis talk introduces the concept of treating humans as asynchronous APIs within agentic systems, emphasizing the need for durable state management when human interaction is required. It demonstrates how Temporal's workers, workflows, and activities can be used to build human-in-the-loop agents that do not block or lose state during human delays, even across worker restarts. The use of 'wait conditions' and 'signals' is highlighted as a mechanism for managing these asynchronous human interactions, preventing production failures caused by blocking calls. The talk also touches upon the decision-making process for when to involve a human in the loop, balancing the cost of errors against alert fatigue.
  5. Your Agents Are in Solitary Confinement: Why MCP & A2A Aren't Enough — Vlad Luzin, BandThis talk argues that current methods for enabling AI agents to work together, such as running models side-by-side and manually routing messages, or using basic scripting, are insufficient. It identifies the 'single-agent bottleneck' and the limitations of human-centric messaging platforms (Slack, Discord) and low-level protocols (MCP, A2A) for bot-to-bot traffic. The core technical contribution is framing agent collaboration as a distributed systems problem, requiring specific architectural components: real-time ordered transport, persistence and rehydration, runtime binding across frameworks, agent-first abstractions (participants, rooms, routing), and enterprise governance. Two demos illustrate agents (Codex, LangGraph, Claude Code, personal assistant) discovering and collaborating in real-time.