Arun Jyoti Chakraborty

I’m an Applied AI Engineer with production experience deploying multi-agent systems, RAG pipelines, and quantitative research agents on AWS and Google Cloud. I focus on low-latency LLM orchestration (MCP/FastMCP, Gemini, Bedrock), agentic workflow design, and robust guardrails.

I’ve built 4 live AI systems with custom domains, automated safety testing, and real-time analytics. My work emphasizes practical reliability—like automated pytest guardrails that verify refusal behavior and block tool calls under adversarial prompt-injection attacks.

Technically, I design decoupled agent architectures so orchestration and tool execution don’t become tightly coupled. For example, I implemented a Gemini orchestrator plus a FastMCP tool server, with persistent session memory via SqliteSessionService and time-series forecasting integrated over HTTP/SSE.

I’m currently pursuing my B.Sc. Data Science & Artificial Intelligence at IIT Guwahati (online, self-paced), and I’m fully available for remote full-time or internship roles immediately. Certified by Microsoft (Azure) and AWS (Bedrock), I’m excited to build trustworthy, scalable AI systems with strong developer ergonomics.