<?xml version="1.0" encoding="UTF-8"?>
<rss xmlns:dc="http://purl.org/dc/elements/1.1/" version="2.0">
  <channel>
    <title>InfoQ - Architecture &amp; Design - Presentations</title>
    <link>https://www.infoq.com</link>
    <description>InfoQ Architecture &amp; Design Presentations feed</description>
    <item>
      <title>Presentation: Scaling LLM-Based Ranking Systems for Latency-Critical Search &amp; Recommendation Workloads</title>
      <link>https://www.infoq.com/presentations/scale-llm-ranking-systems-linkedin/?utm_campaign=infoq_content&amp;utm_source=infoq&amp;utm_medium=feed&amp;utm_term=Architecture+%26+Design-presentations</link>
      <description>&lt;img src="https://res.infoq.com/presentations/scale-llm-ranking-systems-linkedin/en/mediumimage/SundaraRamanRamachandran-medium-1793279019345.jpeg"/&gt;&lt;p&gt;Sundara Raman Ramachandran shares how LinkedIn productionizes LLM ranking under strict P99 latency SLAs. He discusses distilling a 7B model down to 400M parameters, compressing job contexts into single-token embeddings for a 10x throughput leap, and eliminating hidden system bottlenecks - from custom flash attention masks to Python GIL workarounds, gc.freeze, and non-blocking CUDA memory copies.&lt;/p&gt; &lt;i&gt;By Sundara Raman Ramachandran&lt;/i&gt;</description>
      <category>Optimization</category>
      <category>GPU</category>
      <category>QCon AI Boston 2026</category>
      <category>Transcripts</category>
      <category>Low Latency</category>
      <category>Rankings</category>
      <category>Large language models</category>
      <category>LinkedIn</category>
      <category>Architecture &amp; Design</category>
      <category>Development</category>
      <category>AI, ML &amp; Data Engineering</category>
      <category>presentation</category>
      <pubDate>Tue, 17 Nov 2026 11:00:00 GMT</pubDate>
      <guid>https://www.infoq.com/presentations/scale-llm-ranking-systems-linkedin/?utm_campaign=infoq_content&amp;utm_source=infoq&amp;utm_medium=feed&amp;utm_term=Architecture+%26+Design-presentations</guid>
      <dc:creator>Sundara Raman Ramachandran</dc:creator>
      <dc:date>2026-11-17T11:00:00Z</dc:date>
      <dc:identifier>/presentations/scale-llm-ranking-systems-linkedin/en</dc:identifier>
    </item>
    <item>
      <title>Presentation: The Influence Toolkit</title>
      <link>https://www.infoq.com/presentations/bbc-software-architecture-influence/?utm_campaign=infoq_content&amp;utm_source=infoq&amp;utm_medium=feed&amp;utm_term=Architecture+%26+Design-presentations</link>
      <description>&lt;img src="https://res.infoq.com/presentations/bbc-software-architecture-influence/en/mediumimage/henry-wilson-medium-1792653480613.jpg"/&gt;&lt;p&gt;Henry Wilson shares actionable tools from the BBC Sounds re-architecture to help senior engineering leaders drive technical strategy. He discusses building buy-in, navigating active resistors, using RFCs and ADRs, and balancing tradeoffs - showing that effective software architecture is fundamentally about communication, relationship-building, and organizational influence.&lt;/p&gt; &lt;i&gt;By Henry Wilson&lt;/i&gt;</description>
      <category>Architecture Decision Records</category>
      <category>Leadership</category>
      <category>Architecture</category>
      <category>The BBC</category>
      <category>Transcripts</category>
      <category>Staff Plus</category>
      <category>Platform Engineering</category>
      <category>QCon London 2026</category>
      <category>Communication</category>
      <category>Culture &amp; Methods</category>
      <category>Architecture &amp; Design</category>
      <category>Development</category>
      <category>presentation</category>
      <pubDate>Fri, 13 Nov 2026 11:00:00 GMT</pubDate>
      <guid>https://www.infoq.com/presentations/bbc-software-architecture-influence/?utm_campaign=infoq_content&amp;utm_source=infoq&amp;utm_medium=feed&amp;utm_term=Architecture+%26+Design-presentations</guid>
      <dc:creator>Henry Wilson</dc:creator>
      <dc:date>2026-11-13T11:00:00Z</dc:date>
      <dc:identifier>/presentations/bbc-software-architecture-influence/en</dc:identifier>
    </item>
    <item>
      <title>Presentation: Opening the Black Box: Critical Explainability for AI Agent Tool Selection with Kiji Inspector</title>
      <link>https://www.infoq.com/presentations/explainable-ai-llm-agent-decisions/?utm_campaign=infoq_content&amp;utm_source=infoq&amp;utm_medium=feed&amp;utm_term=Architecture+%26+Design-presentations</link>
      <description>&lt;img src="https://res.infoq.com/presentations/explainable-ai-llm-agent-decisions/en/mediumimage/HannesHapke-medium-1793278895069.jpeg"/&gt;&lt;p&gt;Hannes Hapke shares how Dataiku’s open-source Kiji Inspector applies Sparse Autoencoders (SAEs) directly to decision tokens to expose why agents pick specific tools. Moving past superficial chain-of-thought narratives, he discusses causal ablation, layer activation extraction via patched vLLM engines, and production monitoring workflows for software architects and senior engineering leaders.&lt;/p&gt; &lt;i&gt;By Hannes Hapke&lt;/i&gt;</description>
      <category>AI Interpretability</category>
      <category>Governance</category>
      <category>Responsibility</category>
      <category>QCon AI Boston 2026</category>
      <category>AI Architecture</category>
      <category>Transcripts</category>
      <category>Explainable AI</category>
      <category>Agents</category>
      <category>Large language models</category>
      <category>Architecture &amp; Design</category>
      <category>Development</category>
      <category>AI, ML &amp; Data Engineering</category>
      <category>presentation</category>
      <pubDate>Thu, 12 Nov 2026 11:00:00 GMT</pubDate>
      <guid>https://www.infoq.com/presentations/explainable-ai-llm-agent-decisions/?utm_campaign=infoq_content&amp;utm_source=infoq&amp;utm_medium=feed&amp;utm_term=Architecture+%26+Design-presentations</guid>
      <dc:creator>Hannes Hapke</dc:creator>
      <dc:date>2026-11-12T11:00:00Z</dc:date>
      <dc:identifier>/presentations/explainable-ai-llm-agent-decisions/en</dc:identifier>
    </item>
    <item>
      <title>Presentation: Uncorking Queueing Bottlenecks with OpenTelemetry</title>
      <link>https://www.infoq.com/presentations/opentelemetry-queue-observability/?utm_campaign=infoq_content&amp;utm_source=infoq&amp;utm_medium=feed&amp;utm_term=Architecture+%26+Design-presentations</link>
      <description>&lt;img src="https://res.infoq.com/presentations/opentelemetry-queue-observability/en/mediumimage/julian-oli-medium-1792652839179.jpg"/&gt;&lt;p&gt;Oli Lane and Julian Wreford share how Gearset replaced guesswork in distributed asynchronous queues with OpenTelemetry tracing. They explain why traditional metrics fall short, how to implement user-centric SLOs based on latency distributions, and share practical debugging tips to help senior developers, software architects, and engineering leaders improve system reliability and team culture.&lt;/p&gt; &lt;i&gt;By Julian Wreford, Oli Lane&lt;/i&gt;</description>
      <category>Queue</category>
      <category>Asynchronous Architecture</category>
      <category>Transcripts</category>
      <category>Reliability</category>
      <category>Observability</category>
      <category>Distributed Tracing</category>
      <category>OpenTelemetry</category>
      <category>QCon London 2026</category>
      <category>Service Level Objectives</category>
      <category>Culture &amp; Methods</category>
      <category>Architecture &amp; Design</category>
      <category>Development</category>
      <category>DevOps</category>
      <category>presentation</category>
      <pubDate>Wed, 11 Nov 2026 11:00:00 GMT</pubDate>
      <guid>https://www.infoq.com/presentations/opentelemetry-queue-observability/?utm_campaign=infoq_content&amp;utm_source=infoq&amp;utm_medium=feed&amp;utm_term=Architecture+%26+Design-presentations</guid>
      <dc:creator>Julian Wreford, Oli Lane</dc:creator>
      <dc:date>2026-11-11T11:00:00Z</dc:date>
      <dc:identifier>/presentations/opentelemetry-queue-observability/en</dc:identifier>
    </item>
    <item>
      <title>Presentation: Batch Intelligence at Scale: Cost-Efficient Multi-Agent LLM Workflows with Built-In Resilience</title>
      <link>https://www.infoq.com/presentations/scaling-multi-agent-llm-batch/?utm_campaign=infoq_content&amp;utm_source=infoq&amp;utm_medium=feed&amp;utm_term=Architecture+%26+Design-presentations</link>
      <description>&lt;img src="https://res.infoq.com/presentations/scaling-multi-agent-llm-batch/en/mediumimage/AdityaMulik-medium-1793278766506.jpeg"/&gt;&lt;p&gt;Aditya Mulik explains how Walmart Global Tech deployed multi-agent LLM workflows across 4,700 stores for nightly batch inventory forecasting. He discusses tackling production hurdles - schema violations, cross-department data mismatches, and network drops - via MCP grounding and loop agents, while slashing cloud inference costs from $100k+ to $10k per month shifting from PTUs to pay-as-you-go.&lt;/p&gt; &lt;i&gt;By Aditya Mulik&lt;/i&gt;</description>
      <category>Machine Learning</category>
      <category>Resilience</category>
      <category>QCon AI Boston 2026</category>
      <category>Model Context Protocol (MCP)</category>
      <category>Transcripts</category>
      <category>Batch Processing</category>
      <category>Agents</category>
      <category>Large language models</category>
      <category>AI Cost Optimisation</category>
      <category>Architecture &amp; Design</category>
      <category>Development</category>
      <category>AI, ML &amp; Data Engineering</category>
      <category>presentation</category>
      <pubDate>Tue, 10 Nov 2026 11:00:00 GMT</pubDate>
      <guid>https://www.infoq.com/presentations/scaling-multi-agent-llm-batch/?utm_campaign=infoq_content&amp;utm_source=infoq&amp;utm_medium=feed&amp;utm_term=Architecture+%26+Design-presentations</guid>
      <dc:creator>Aditya Mulik</dc:creator>
      <dc:date>2026-11-10T11:00:00Z</dc:date>
      <dc:identifier>/presentations/scaling-multi-agent-llm-batch/en</dc:identifier>
    </item>
  </channel>
</rss>
