Aneesh Amaravadi

I am a Computer Science and Engineering student at CBIT focused on building resilient multi-agent systems and stateful orchestration runtimes. I specialize in designing local-first AI architectures that leverage durable execution, human-in-the-loop (HITL) evaluator loops, and robust persistent memory to transition agentic workflows from experimental wrappers into production-ready pipelines. My recent work centers on complex system architecture and predictive analytics. I developed AgentForge Local, a research OS and execution system built with Python and LangGraph that handles pause/resume checkpoints and node-level execution traces. I also architected PAIMANA AI, a multi-agent early-warning decision-support system, and built IncidentIQ, an AI production incident response memory agent using Hindsight persistent memory. Beyond AI orchestration, I maintain a strong focus on core backend engineering, concurrency, and data structures. I built MiniRedis, a concurrent, custom TCP-based in-memory data store developed entirely from scratch in Java to manage bounded multi-client concurrency and atomic operations. I am actively building secure, scalable agentic pipelines utilizing MCP gateways, hard runtimes, and local knowledge graphs