<?xml version="1.0" encoding="UTF-8"?>
<rss version="2.0" xmlns:atom="http://www.w3.org/2005/Atom">
  <channel>
    <title>Paurush Muley | Engineering Blog &amp; Systems Research</title>
    <link>https://www.paurushmuley.com/blog</link>
    <description>Deep-dive technical articles on .NET 10, Agentic AI, LangGraph, Microservices, and Cloud Systems by Paurush Muley.</description>
    <language>en-us</language>
    <lastBuildDate>Wed, 21 Oct 2026 07:36:28 GMT</lastBuildDate>
    <atom:link href="https://www.paurushmuley.com/rss.xml" rel="self" type="application/rss+xml"/>
    <item>
      <title>Designing Multi-Region Active-Active API Gateways with Envoy, Istio, and Azure Front Door in .NET 10</title>
      <link>https://www.paurushmuley.com/blog/multi-region-active-active-api-gateways-envoy-dotnet-10</link>
      <guid isPermaLink="true">https://www.paurushmuley.com/blog/multi-region-active-active-api-gateways-envoy-dotnet-10</guid>
      <description>An enterprise engineering guide to architecting resilient, multi-region active-active API gateway topologies in .NET 10. Learn how to combine Azure Front Door Anycast edge routing, Envoy proxies with distributed token-bucket rate limiting, and multi-cluster Istio mTLS to eliminate DNS failover lag, prevent split-brain dual writes, and achieve sub-second global failover.</description>
      <pubDate>Wed, 21 Oct 2026 00:00:00 GMT</pubDate>
      <author>paurushmuley@gmail.com (Paurush Muley)</author>
      <category>Envoy</category><category>Istio</category><category>Azure Front Door</category><category>Multi-Region</category><category>.NET 10</category><category>API Gateway</category><category>Cloud Architecture</category><category>Microservices</category><category>High Availability</category><category>C#</category><category>Kubernetes</category>
      <enclosure url="https://bcefknifzbdns0qe.public.blob.vercel-storage.com/blogs/multi-region-active-active-api-gateways-envoy-dotnet-10.jpg" length="0" type="image/jpeg"/>
    </item>
    <item>
      <title>Observability at Scale: OpenTelemetry, Distributed Tracing, and eBPF in Cloud-Native .NET 10 Microservices</title>
      <link>https://www.paurushmuley.com/blog/observability-at-scale-opentelemetry-ebpf-dotnet-10</link>
      <guid isPermaLink="true">https://www.paurushmuley.com/blog/observability-at-scale-opentelemetry-ebpf-dotnet-10</guid>
      <description>Architect zero-allocation distributed tracing and kernel-level network observability in high-concurrency .NET 10 microservices. Learn how to eliminate APM CPU profiling penalties, propagate W3C Trace Context across Apache Kafka, enforce tail-based sampling with OpenTelemetry Collectors, and combine eBPF telemetry for end-to-end visibility.</description>
      <pubDate>Mon, 19 Oct 2026 00:00:00 GMT</pubDate>
      <author>paurushmuley@gmail.com (Paurush Muley)</author>
      <category>OpenTelemetry</category><category>Observability</category><category>Distributed Tracing</category><category>eBPF</category><category>.NET 10</category><category>Microservices</category><category>Azure Monitor</category><category>Performance</category><category>Kafka</category><category>C#</category><category>Cloud Native</category>
      <enclosure url="https://bcefknifzbdns0qe.public.blob.vercel-storage.com/blogs/observability-at-scale-opentelemetry-ebpf-dotnet-10.jpg" length="0" type="image/jpeg"/>
    </item>
    <item>
      <title>Enterprise Semantic Caching with Redis &amp; Vector Embeddings: Slashing LLM Costs by 55% in .NET Microservices</title>
      <link>https://www.paurushmuley.com/blog/enterprise-semantic-caching-redis-vector-embeddings-dotnet</link>
      <guid isPermaLink="true">https://www.paurushmuley.com/blog/enterprise-semantic-caching-redis-vector-embeddings-dotnet</guid>
      <description>A production engineering deep dive into implementing high-performance semantic caching in .NET 10 microservices using Redis vector similarity search (FT.SEARCH with HNSW) and lightweight embeddings. Slashing LLM token costs by over 55%, reducing p95 latency from 1,200ms to 14ms, and eliminating cache stampedes with two-tier L1/L2 orchestration.</description>
      <pubDate>Mon, 19 Oct 2026 00:00:00 GMT</pubDate>
      <author>paurushmuley@gmail.com (Paurush Muley)</author>
      <category>Semantic Caching</category><category>Redis</category><category>Vector Embeddings</category><category>.NET 10</category><category>LLM Gateway</category><category>Cost Optimization</category><category>Azure OpenAI</category><category>Microservices</category>
      <enclosure url="https://bcefknifzbdns0qe.public.blob.vercel-storage.com/blogs/enterprise-semantic-caching-redis-vector-embeddings-dotnet.jpg" length="0" type="image/jpeg"/>
    </item>
    <item>
      <title>Deploying Local Small Language Models (SLMs) with ONNX Runtime &amp; C# in Containerized Cloud Microservices</title>
      <link>https://www.paurushmuley.com/blog/local-slms-onnx-runtime-csharp-microservices</link>
      <guid isPermaLink="true">https://www.paurushmuley.com/blog/local-slms-onnx-runtime-csharp-microservices</guid>
      <description>A production engineering guide to hosting quantized Small Language Models (Phi-4-mini) directly inside .NET 10 microservice containers using ONNX Runtime GenAI. Slashing inference latency to sub-25ms, bypassing cloud API rate limits, and eliminating egress costs with native tensor pooling.</description>
      <pubDate>Fri, 16 Oct 2026 00:00:00 GMT</pubDate>
      <author>paurushmuley@gmail.com (Paurush Muley)</author>
      <category>SLM</category><category>ONNX Runtime</category><category>.NET 10</category><category>C#</category><category>Local AI</category><category>Cloud Microservices</category><category>Phi-4</category><category>Edge Computing</category>
      <enclosure url="https://bcefknifzbdns0qe.public.blob.vercel-storage.com/blogs/local-slms-onnx-runtime-csharp-microservices.jpg" length="0" type="image/jpeg"/>
    </item>
    <item>
      <title>Distributed Transaction Sagas with Kafka &amp; Azure Service Bus in .NET 10: Solving Dual-Write Hazards with the Outbox Pattern</title>
      <link>https://www.paurushmuley.com/blog/distributed-transaction-sagas-kafka-azure-dotnet-10</link>
      <guid isPermaLink="true">https://www.paurushmuley.com/blog/distributed-transaction-sagas-kafka-azure-dotnet-10</guid>
      <description>A production deep dive into solving dual-write hazards and executing resilient distributed transaction sagas in .NET 10 using EF Core transactional outboxes, high-throughput bounded Channels, Apache Kafka, and Azure Service Bus with idempotent compensating flows.</description>
      <pubDate>Fri, 16 Oct 2026 00:00:00 GMT</pubDate>
      <author>paurushmuley@gmail.com (Paurush Muley)</author>
      <category>.NET 10</category><category>Distributed Sagas</category><category>Kafka</category><category>Azure Service Bus</category><category>Outbox Pattern</category><category>Microservices</category><category>CQRS</category><category>Event-Driven Architecture</category><category>C#</category><category>System.Threading.Channels</category>
      <enclosure url="https://bcefknifzbdns0qe.public.blob.vercel-storage.com/blogs/distributed-transaction-sagas-kafka-azure-dotnet-10.jpg" length="0" type="image/jpeg"/>
    </item>
    <item>
      <title>Zero-Allocation Event Processing in .NET 10: Building 100k+ RPS Stream Processors with System.IO.Pipelines &amp; Bounded Channels</title>
      <link>https://www.paurushmuley.com/blog/zero-allocation-event-processing-dotnet-10-pipelines</link>
      <guid isPermaLink="true">https://www.paurushmuley.com/blog/zero-allocation-event-processing-dotnet-10-pipelines</guid>
      <description>A production deep dive into building ultra-low-latency, zero-allocation event stream processors in .NET 10 using System.IO.Pipelines, SequenceReader&lt;byte&gt;, and bounded Channel&lt;T&gt; backpressure to eliminate GC pauses and sustain 100k+ RPS under heavy enterprise workloads.</description>
      <pubDate>Thu, 15 Oct 2026 00:00:00 GMT</pubDate>
      <author>paurushmuley@gmail.com (Paurush Muley)</author>
      <category>.NET 10</category><category>High Throughput</category><category>System.IO.Pipelines</category><category>Memory Management</category><category>Kafka</category><category>Microservices</category><category>Span&lt;T&gt;</category><category>C#</category>
      <enclosure url="https://bcefknifzbdns0qe.public.blob.vercel-storage.com/blogs/zero-allocation-event-processing-dotnet-10-pipelines.jpg" length="0" type="image/jpeg"/>
    </item>
    <item>
      <title>High-Throughput Microservice Resiliency in .NET 10: Polly v8 Pipelines &amp; Distributed Outbox Patterns</title>
      <link>https://www.paurushmuley.com/blog/microservice-resiliency-dotnet-10-polly-v8</link>
      <guid isPermaLink="true">https://www.paurushmuley.com/blog/microservice-resiliency-dotnet-10-polly-v8</guid>
      <description>A production deep dive into engineering fault-tolerant enterprise microservices in .NET 10 using Polly v8 non-allocating resilience pipelines, transactional outbox daemons, and sub-10ms circuit breakers.</description>
      <pubDate>Thu, 15 Oct 2026 00:00:00 GMT</pubDate>
      <author>paurushmuley@gmail.com (Paurush Muley)</author>
      <category>.NET 10</category><category>Polly v8</category><category>Microservices</category><category>Kafka</category><category>Resilience</category><category>Azure</category><category>High Concurrency</category>
      <enclosure url="https://bcefknifzbdns0qe.public.blob.vercel-storage.com/blogs/microservice-resiliency-dotnet-10-polly-v8-1792041140211.jpg" length="0" type="image/jpeg"/>
    </item>
    <item>
      <title>High-Performance Hybrid Vector Search in Enterprise RAG: Implementing Reciprocal Rank Fusion &amp; Cross-Encoder Reranking in Azure</title>
      <link>https://www.paurushmuley.com/blog/hybrid-vector-search-reciprocal-rank-fusion-rag</link>
      <guid isPermaLink="true">https://www.paurushmuley.com/blog/hybrid-vector-search-reciprocal-rank-fusion-rag</guid>
      <description>A production engineering blueprint for building high-performance enterprise hybrid search in .NET 10 and Azure, fusing BM25 lexical precision with dense vector embeddings via Reciprocal Rank Fusion (RRF) and ONNX Cross-Encoder reranking.</description>
      <pubDate>Thu, 15 Oct 2026 00:00:00 GMT</pubDate>
      <author>paurushmuley@gmail.com (Paurush Muley)</author>
      <category>Vector Search</category><category>RAG</category><category>Reciprocal Rank Fusion</category><category>Azure OpenAI</category><category>GraphRAG</category><category>Agentic AI</category><category>BM25</category><category>Enterprise Search</category><category>.NET 10</category><category>C#</category>
      <enclosure url="https://bcefknifzbdns0qe.public.blob.vercel-storage.com/blogs/hybrid-vector-search-reciprocal-rank-fusion-rag.jpg" length="0" type="image/jpeg"/>
    </item>
    <item>
      <title>Engineering Multi-Agent Tool Guardrails in Distributed .NET Microservices: Securing Autonomous Execution Beyond System Prompts</title>
      <link>https://www.paurushmuley.com/blog/multi-agent-tool-guardrails-distributed-dotnet-microservices</link>
      <guid isPermaLink="true">https://www.paurushmuley.com/blog/multi-agent-tool-guardrails-distributed-dotnet-microservices</guid>
      <description>A deep architectural blueprint for engineering deterministic, zero-trust tool execution guardrails in distributed .NET 10 microservices using Microsoft.Extensions.AI, attribute-based policy filters, Polly circuit breakers, and sub-millisecond memory-safe sandboxing.</description>
      <pubDate>Wed, 14 Oct 2026 00:00:00 GMT</pubDate>
      <author>paurushmuley@gmail.com (Paurush Muley)</author>
      <category>Agentic AI</category><category>Tool Guardrails</category><category>.NET 10</category><category>Microservices Security</category><category>Distributed Systems</category><category>Semantic Kernel</category><category>Zero Trust</category>
      <enclosure url="https://bcefknifzbdns0qe.public.blob.vercel-storage.com/blogs/multi-agent-tool-guardrails-distributed-dotnet-microservices.jpg" length="0" type="image/jpeg"/>
    </item>
    <item>
      <title>Architecting Autonomous Multi-Agent Workflows in .NET 10: Combining C# Type Safety with LangGraph &amp; Native AOT</title>
      <link>https://www.paurushmuley.com/blog/architecting-autonomous-multi-agent-workflows-dotnet-10</link>
      <guid isPermaLink="true">https://www.paurushmuley.com/blog/architecting-autonomous-multi-agent-workflows-dotnet-10</guid>
      <description>A comprehensive architectural guide to engineering high-scale, cyclic multi-agent graphs in .NET 10 using Microsoft.Extensions.AI and C# type safety, compiled with Native AOT for sub-15ms cold starts in cloud-native microservices.</description>
      <pubDate>Wed, 14 Oct 2026 00:00:00 GMT</pubDate>
      <author>paurushmuley@gmail.com (Paurush Muley)</author>
      <category>.NET 10</category><category>Agentic AI</category><category>C#</category><category>LangGraph</category><category>Native AOT</category><category>Azure Microservices</category><category>Semantic Kernel</category>
      <enclosure url="https://bcefknifzbdns0qe.public.blob.vercel-storage.com/blogs/architecting-autonomous-multi-agent-workflows-dotnet-10.jpg" length="0" type="image/jpeg"/>
    </item>
    <item>
      <title>Building Production-Ready Agentic RAG Systems with LangChain &amp; Azure</title>
      <link>https://www.paurushmuley.com/blog/building-agentic-rag-pipelines</link>
      <guid isPermaLink="true">https://www.paurushmuley.com/blog/building-agentic-rag-pipelines</guid>
      <description>A deep dive into transitioning from naive vector search RAG to autonomous agentic workflows that self-correct, validate queries, and execute multi-hop reasoning in enterprise environments.</description>
      <pubDate>Tue, 12 Nov 2024 00:00:00 GMT</pubDate>
      <author>paurushmuley@gmail.com (Paurush Muley)</author>
      <category>Agentic AI</category><category>RAG</category><category>LangChain</category><category>LangGraph</category><category>Azure</category><category>LLMs</category>
      <enclosure url="https://bcefknifzbdns0qe.public.blob.vercel-storage.com/blogs/building-agentic-rag-pipelines.jpg" length="0" type="image/jpeg"/>
    </item>
    <item>
      <title>Optimizing .NET Core Microservices for High-Concurrency Cloud Workflows</title>
      <link>https://www.paurushmuley.com/blog/high-throughput-microservices-dotnet</link>
      <guid isPermaLink="true">https://www.paurushmuley.com/blog/high-throughput-microservices-dotnet</guid>
      <description>Techniques for optimizing .NET Core REST APIs, async event loops, connection pooling, and Azure deployments to achieve a 30% throughput improvement.</description>
      <pubDate>Fri, 20 Sep 2024 00:00:00 GMT</pubDate>
      <author>paurushmuley@gmail.com (Paurush Muley)</author>
      <category>.NET Core</category><category>Microservices</category><category>Azure</category><category>Kafka</category><category>Performance</category><category>Cloud Architecture</category>
      <enclosure url="https://bcefknifzbdns0qe.public.blob.vercel-storage.com/blogs/high-throughput-microservices-dotnet.jpg" length="0" type="image/jpeg"/>
    </item>
    <item>
      <title>Bridging the Gap: Integrating LLMs Seamlessly into Modern React Frontends</title>
      <link>https://www.paurushmuley.com/blog/modern-fullstack-ai-integration</link>
      <guid isPermaLink="true">https://www.paurushmuley.com/blog/modern-fullstack-ai-integration</guid>
      <description>How to design responsive, streaming AI interfaces in React with typewriter animations, optimistic UI updates, grounding citations, and recruiter-friendly UX.</description>
      <pubDate>Mon, 05 Aug 2024 00:00:00 GMT</pubDate>
      <author>paurushmuley@gmail.com (Paurush Muley)</author>
      <category>React</category><category>Next.js</category><category>TypeScript</category><category>UX Design</category><category>Generative AI</category><category>Full-Stack</category>
      <enclosure url="https://bcefknifzbdns0qe.public.blob.vercel-storage.com/blogs/modern-fullstack-ai-integration.jpg" length="0" type="image/jpeg"/>
    </item>
  </channel>
</rss>