What Agent Substrate on GKE made me ask about workers, checkpoints, and tool calls.
Google's Agent Substrate announcement is a density story. I kept asking what actually gets frozen, and whether a restored agent would replay a tool call that already succeeded.
We live in an age where misinformation travels faster than the truth. False claims ripple through social media and news aggregators, often outpacing the human ability to fact-check them. This isn't just annoying; it poses severe risks to public trust and decision-making.As a researcher and engineer, I wanted to tackle this head-on. I didn't just want another text classifier; I wanted a system that could 'think' through a lie.
If the Backend is the muscle and AWS is the nervous system, then the AI Agent Layer is the brain of NewsInsight. When I started this project, I didn't just want to display news. I wanted to understand it. I wanted to know if an article was happy or sad, if it was about 'Apple the company' or 'apple the fruit,' and most importantly, if it was actually news or just spam. Here is how I architected the AI pipeline using AWS Bedrock and Claude 3.5 Haiku.
I can't stay awake 24/7 to curate news, so I built a system that does it for me. The real magic of NewsInsight happens while I'm asleep. This is a deep dive into the AWS Serverless Architecture that powers the project. I’ll explain how I orchestrated a 'Night Shift' using EventBridge and Glue to harvest news automatically, and how I integrated AWS Bedrock to read and summarize thousands of articles before my alarm goes off. If you want to see how to build a 'set it and forget it' data pipeline, this is the post for you.
A beautiful interface is useless if it’s waiting 10 seconds for data. While the frontend is the face of NewsInsight, the Backend is the brain that makes it smart. This article dives into the engine room. I’ll share how I architected the FastAPI backend to handle search requests, built a custom 'Smart Caching' layer to keep costs low, and designed the logic that connects my user interface to the raw power of AWS. This is the story of how I made the app fast, resilient, and intelligent.
We live in an age of information overload. My original spark for NewsInsight was simple: I was tired of doom-scrolling. I wanted a dashboard that didn't just dump links on me, but actually helped me visualize what was happening in the world. In this post, I’ll walk you through how I built the Frontend. I’ll explain why I chose React and Tailwind to create a distraction-free interface, how I visualized AI sentiment data, and the challenges of making a complex data app feel simple to the user.
An introduction to Java threads, the memory model, race conditions, and how synchronization protects shared state.
Understanding Docker, Postgres, GCP, and Terraform with respect to Data Engineering
AWS Compute Services provide a robust suite of solutions to meet diverse computing needs, from virtual servers with Amazon EC2 to serverless execution with AWS Lambda. Whether you're running containerized applications with ECS and EKS, or simplifying deployments with AWS Fargate, AWS offers the flexibility to scale workloads efficiently while optimizing costs. Dive into how these services empower businesses to innovate and grow with unparalleled performance and reliability.
Explore the concepts behind core AWS services