Yash Chudasama
Real-Time Communication · AI · Cloud
I architect systems where milliseconds matter — WebRTC, voice AI pipelines, and cloud infrastructure at scale. I write about the hard problems in real-time communication, AI, and distributed systems.
Areas of Expertise
Real-Time Communication
WebRTC, SFU/MCU architecture, media pipelines, low-latency protocols, and scaling live audio & video.
AI & Intelligence
Voice AI, LLMs in production, RAG systems, AI agents, and real-time intelligence pipelines.
Cloud Infrastructure
AWS, GCP, Azure at scale. Kubernetes, edge computing, and distributed systems for latency-sensitive workloads.
Featured Writing
SFU vs MCU: Choosing the Right Video Architecture
When to use a Selective Forwarding Unit vs a Multipoint Control Unit for WebRTC — bandwidth, latency, cost, and scaling trade-offs from production experience
SFU vs MCU: Choosing the Right Video ArchitectureVoice AI Latency: Why Sub-200ms Response Time Matters
Engineering sub-200ms voice AI responses — STT, LLM, and TTS pipeline optimization, streaming architecture, and latency trade-offs
Voice AI Latency: Why Sub-200ms Response Time MattersAI Agents: The Rise of Autonomous Systems
How AI agents are transforming software through autonomous decision-making and multi-agent collaboration
AI Agents: The Rise of Autonomous SystemsEdge Computing: Bringing Intelligence Closer to Users
Why edge computing is reshaping cloud architecture, reducing latency, and enabling real-time AI applications at scale
Edge Computing: Bringing Intelligence Closer to UsersLatest
Cloud Architecture for Real-Time Media Workloads
Why standard cloud patterns fail for WebRTC and media workloads — instance selection, network topology, auto-scaling pitfalls, and the multi-cloud reality
Cloud Architecture for Real-Time Media WorkloadsReal-Time Voice AI Pipelines: STT, LLM, and TTS
How to architect AI-powered voice pipelines within WebRTC infrastructure — from microphone to intelligent response in under 500ms
Real-Time Voice AI Pipelines: STT, LLM, and TTSWebRTC Debugging: Production Problems and Fixes
NAT traversal failures, one-way audio, codec negotiation edge cases, and media server scaling — hard lessons from building real-time communication systems
WebRTC Debugging: Production Problems and FixesAI and Human Creativity: Collaboration, Not Replacement
Exploring why artificial intelligence amplifies human creativity rather than replacing it, and what this means for the future of work and art
AI and Human Creativity: Collaboration, Not ReplacementWebAssembly Beyond the Browser: The Universal Runtime
How WebAssembly is evolving from a browser technology into a universal runtime for servers, edge computing, IoT, and plugin systems
WebAssembly Beyond the Browser: The Universal RuntimeRAG: Building AI Systems That Know Your Data
A practical guide to Retrieval-Augmented Generation — how to build AI applications that leverage your own data for accurate, grounded responses
RAG: Building AI Systems That Know Your Data