<?xml version="1.0" encoding="UTF-8"?><rss xmlns:dc="http://purl.org/dc/elements/1.1/" xmlns:content="http://purl.org/rss/1.0/modules/content/" xmlns:atom="http://www.w3.org/2005/Atom" version="2.0" xmlns:cc="http://cyber.law.harvard.edu/rss/creativeCommonsRssModule.html">
    <channel>
        <title><![CDATA[Stories by Lance Gabriel on Medium]]></title>
        <description><![CDATA[Stories by Lance Gabriel on Medium]]></description>
        <link>https://medium.com/@lancegabrielwork91?source=rss-44d55fc8d964------2</link>
        <image>
            <url>https://cdn-images-1.medium.com/fit/c/150/150/1*o1iq8AQa6havOJFXNTRjyA.jpeg</url>
            <title>Stories by Lance Gabriel on Medium</title>
            <link>https://medium.com/@lancegabrielwork91?source=rss-44d55fc8d964------2</link>
        </image>
        <generator>Medium</generator>
        <lastBuildDate>Fri, 20 Nov 2026 13:55:08 GMT</lastBuildDate>
        <atom:link href="https://medium.com/@lancegabrielwork91/feed" rel="self" type="application/rss+xml"/>
        <webMaster><![CDATA[yourfriends@medium.com]]></webMaster>
        <atom:link href="http://medium.superfeedr.com/" rel="hub"/>
        <item>
            <title><![CDATA[Instant answers in the browser with Edge Copilot]]></title>
            <link>https://medium.com/@lancegabrielwork91/instant-answers-in-the-browser-with-edge-copilot-d97f739705f2?source=rss-44d55fc8d964------2</link>
            <guid isPermaLink="false">https://medium.com/p/d97f739705f2</guid>
            <category><![CDATA[programming]]></category>
            <category><![CDATA[microsoft-copilot]]></category>
            <category><![CDATA[ai]]></category>
            <dc:creator><![CDATA[Lance Gabriel]]></dc:creator>
            <pubDate>Sat, 11 Oct 2025 05:56:54 GMT</pubDate>
            <atom:updated>2025-10-11T06:01:06.238Z</atom:updated>
            <content:encoded><![CDATA[<p>I’m probably late to this, but after switching from Chrome to Microsoft Edge and updating, I noticed a small feature that’s made a big difference. Press <strong>Ctrl + Shift + .</strong> and Copilot pops open in a side panel. It can answer questions about whatever you’re looking at articles, docs, even videos without leaving the page.</p><p>Since finding it, I’ve been using Copilot a lot. If I’m reading something long, I’ll ask for a quick summary before deciding whether to dive in. If I hit an unfamiliar term, I ask for a definition in context. When I’m watching a video and someone mentions a tool or app, I just tap the shortcut and ask what it is.</p><p>A recent example: I was watching a clip where someone casually referenced a cosplay app. I pressed <strong>Ctrl + Shift + .</strong>, asked “How does the girl in the video make money in the app she was using?” and Copilot pulled up the answers within seconds. Even her name. No new tabs and no need to pause the video.</p><p>The best part is that it keeps me in flow. I can get just enough information to move forward, then get back to whatever I was doing. If you’re on Edge, try the shortcut. It makes the browser feel a little bit like a second brain. If you use Chrome or other browsers, you might have to look around to find a similar feature.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*IcD7VseDmcJVRSZmqOkX9w.png" /></figure><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=d97f739705f2" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Learning AWS Bedrock and LangGraph: My First Steps into Agent Orchestration]]></title>
            <link>https://medium.com/@lancegabrielwork91/learning-aws-bedrock-and-langgraph-my-first-steps-into-agent-orchestration-ec0655068c2e?source=rss-44d55fc8d964------2</link>
            <guid isPermaLink="false">https://medium.com/p/ec0655068c2e</guid>
            <category><![CDATA[langgraph]]></category>
            <category><![CDATA[coding]]></category>
            <category><![CDATA[amazon-bedrock]]></category>
            <category><![CDATA[technology]]></category>
            <category><![CDATA[ai]]></category>
            <dc:creator><![CDATA[Lance Gabriel]]></dc:creator>
            <pubDate>Sat, 11 Oct 2025 02:40:42 GMT</pubDate>
            <atom:updated>2025-10-11T05:00:03.648Z</atom:updated>
            <content:encoded><![CDATA[<h3>My First Post</h3><p>Hi everyone! This is my very first post!</p><p>Through this blog, I plan to share my thoughts and experiences as I build, learn, and occasionally break things in tech. Whenever I run into problems and manage to solve them, I’ll document them here. Hopefully it’ll help out fellow developers who face the same issues.</p><h3>Getting Back Into the Groove</h3><p>For almost a year since I left corporate working on mostly government projects, I’ve been freelancing, mostly building <strong>mobile apps with React Native and Firebase</strong>, and occasionally working on <strong>web apps using Next.js</strong>. It’s been a great journey working on different projects, but lately I’ve been itching to explore something new and a little more futuristic.</p><p>So, I decided to dive into the world of <strong>AI agents</strong> and <strong>agentic workflows, </strong>building systems that can reason, plan, and automate real-world tasks on their own. To challenge myself and stay up to date with the latest tech trends, I kicked things off with a small “pet” <strong>micro-SaaS project</strong> powered by AI agents.</p><p>After researching the current tech landscape, I decided to go all in with <strong>AWS Bedrock</strong>, mainly because it lets me invoke different AI models depending on each agent’s use case.</p><p>I won’t go into full detail about the project’s features just yet. But in short, this project is both a playground for sharpening my skills and a serious portfolio piece. My main goal is to solve real pain points for small business owners, especially solo entrepreneurs who want to save costs through automation.</p><h3>Tech Stack &amp; Architecture</h3><p>Here’s a quick overview of the project’s architecture:</p><ul><li><strong>Frontend:</strong> Deployed on <strong>Vercel</strong> using <strong>Next.js</strong> and <strong>Shadcn UI</strong> for a clean, modern component library and design system</li><li><strong>Backend / AI Orchestration:</strong> Hosted on <strong>AWS</strong>, leveraging <strong>Bedrock</strong>, <strong>LangGraph</strong>, <strong>Lambda</strong>, <strong>DynamoDB</strong>, and <strong>EventBridge</strong></li><li><strong>Models &amp; Embeddings:</strong> Using <strong>Claude 3.5 Haiku</strong> for most text-related reasoning tasks (and occasionally <strong>Claude 3 Haiku</strong> for simpler or lightweight operations since it’s more cost-efficient), while <strong>Amazon Titan Embeddings</strong> handle the vectorization of text data</li><li><strong>Vector Database:</strong> <strong>Qdrant</strong>, for storing and retrieving embeddings efficiently during agent context retrieval</li></ul><p>Getting the setup right took the most time even with ChatGPT’s help. I’m completely new to AWS and didn’t want to just follow instructions blindly. I wanted to <em>understand</em> what I was doing so I took my time exploring the AWS console, figuring things out piece by piece.</p><p>Bonus: I also got <strong>$200 in free AWS credits</strong>, which made experimenting a lot more fun.</p><h3>First Impressions of AWS Bedrock</h3><p>For those unfamiliar, <strong>Amazon Bedrock</strong> is AWS’s suite of tools for building AI applications. You can use it to access and invoke various foundation models directly within your AWS environment.</p><p>Through this project (still a work in progress), AWS doesn’t feel as intimidating anymore, it’s much less foreign compared to how it felt back then.</p><h3>Hitting Roadblocks</h3><p>One issue I ran into: when I tried to call an AI model from my Next.js app using the AWS CLI, I kept getting a “permission denied” error, even though I had already enabled model access in the Bedrock console.</p><p>After some troubleshooting, I realized there were <strong>two key issues</strong>:</p><ol><li>You need to <strong>switch your AWS region</strong> in the top-right corner to enable access for each specific model individually (a surprisingly easy thing to miss).</li><li><strong>Claude 3.5 models require an Inference Profile</strong> in Bedrock before you can actually invoke them, without that setup, even if access is enabled, you’ll keep running into permission errors.</li></ol><p>That first one took me a while to figure out, since the documentation wasn’t entirely clear or maybe I did not search hard enough. Honestly, it’s a bit of a UX flaw but AWS mentioned that these steps will soon be simplified, so I’m looking forward to that update.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/442/0*K0BbOOfv0OGi8e9l" /></figure><h3>Wrapping Up</h3><p>That’s it for now! I’ll continue building and documenting my progress as I go. This project has already been a great learning experience both technically and personally.</p><p>Time to get back to coding.</p><p>P.S. If you have any good sources or platforms to stay updated on the latest tech trends, feel free to drop a comment or email me</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=ec0655068c2e" width="1" height="1" alt="">]]></content:encoded>
        </item>
    </channel>
</rss>