<?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 Sarath on Medium]]></title>
        <description><![CDATA[Stories by Sarath on Medium]]></description>
        <link>https://medium.com/@sarath.hotspot?source=rss-c828a6bcd49a------2</link>
        <image>
            <url>https://cdn-images-1.medium.com/fit/c/150/150/0*96zkisrJSk2yge_z</url>
            <title>Stories by Sarath on Medium</title>
            <link>https://medium.com/@sarath.hotspot?source=rss-c828a6bcd49a------2</link>
        </image>
        <generator>Medium</generator>
        <lastBuildDate>Thu, 19 Nov 2026 11:30:57 GMT</lastBuildDate>
        <atom:link href="https://medium.com/@sarath.hotspot/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[Your App Idea Deserves to Be Built — Here’s How (No Coding Needed) with Flutterflow.]]></title>
            <link>https://medium.com/@sarath.hotspot/your-app-idea-deserves-to-be-built-heres-how-no-coding-needed-with-flutterflow-55ebf26968c4?source=rss-c828a6bcd49a------2</link>
            <guid isPermaLink="false">https://medium.com/p/55ebf26968c4</guid>
            <category><![CDATA[no-code]]></category>
            <dc:creator><![CDATA[Sarath]]></dc:creator>
            <pubDate>Sun, 09 Feb 2025 07:07:35 GMT</pubDate>
            <atom:updated>2025-02-09T07:07:35.987Z</atom:updated>
            <content:encoded><![CDATA[<h3>Your App Idea Deserves to Be Built — Here’s How (No Coding Needed) with Flutterflow.</h3><p>I recently came across Flutterflow, a No-Code, full fledged app/website building tool. Since then it became a powerful tool in my app building journey.</p><p>I have built multiple apps which have ~100K active users.</p><p><a href="https://itunes.apple.com/in/app/india-gold-silver-mcx-prices/id963424638">India Gold Silver Price</a> &amp; <a href="https://apps.apple.com/us/app/dad-joke-ai/id6736678054">Dad Joke AI</a></p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/0*zVqg64TUWjPPQkR8" /></figure><h3>1) App idea to reality</h3><p>Start with rough sketches, flutter flow provides a powerfull drag and drop UI tools to build your UI.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/0*nTUEA-gu7cleR6HZ" /></figure><h4>1.1) Write logic also with drag and drop building blocks</h4><p>Write your actions like, button actions, data fetches using drag and drop actions. You can build action chains with visual chains.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/838/1*Xcn1WUd_sM-2qunpq1ouUA.png" /></figure><h4>1.2) Integration with firebase for functionality</h4><p>Flutterflow supports in-built integration with firebase. All necessary app functions like authentication(email, google auth etc), database, cloud storage, push notifications are just clicks away.</p><h4>1.3) Seamless mobile deployment</h4><p>Flutterflow support direct mobile deployment. Once you have your UI ready, you can deploy to Appstore or playstore with in minutes.</p><h3>2) Dad Joke AI</h3><p>My second app “<a href="https://dadjokeai.com/">Dad Joke AI</a>” is released recently and I am able to build the UI in a matter of weeks. Given that I am new to IPhone UI development, for me deploying an app in 7days is like a magic. On top of that, I have single codebase and can deploy same app to both Appstore and playstore.</p><p>Dad Joke AI, is born from curiosity of a dad jokes fan. I kept reading lot of dad jokes and I wonder if we present the joke in visual form, does it enhance the freshness of the joke?</p><p>I built a proto type using flutterflow with in a week and deployed it immediately.</p><p><a href="https://apps.apple.com/us/app/dad-joke-ai/id6736678054">Give it a try and provide your feedback.</a></p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=55ebf26968c4" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Wrappers — Boring yet potential opportunities in the new LLM world!]]></title>
            <link>https://medium.com/@sarath.hotspot/wrappers-boring-yet-potential-opportunities-in-the-new-llm-world-17ca29edc92d?source=rss-c828a6bcd49a------2</link>
            <guid isPermaLink="false">https://medium.com/p/17ca29edc92d</guid>
            <category><![CDATA[llmops]]></category>
            <category><![CDATA[ai]]></category>
            <category><![CDATA[startup]]></category>
            <category><![CDATA[llm-applications]]></category>
            <dc:creator><![CDATA[Sarath]]></dc:creator>
            <pubDate>Mon, 27 Jan 2025 06:36:11 GMT</pubDate>
            <atom:updated>2025-01-27T06:36:11.929Z</atom:updated>
            <content:encoded><![CDATA[<h3>Wrappers — Boring yet potential opportunities in the new LLM world!</h3><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*F28yR8pbb7Dl6R-0HahlIQ.png" /></figure><p>You are just wrapper over Open AI, how do you provide value in LLM world. This is the common question to all startups in AI space. We wonder does wrappers really add value or do they really have moat. I think wrappers handling cross cutting concerns add business value.</p><p>There are new advancements in models every month making headlines, the focus will be on real-world applications. As these use cases gain prominence, they bring a set of cross cutting concerns. E.g. During early days of mobile app development, every app developer need to build their own tools to handle, telemetry, crash reporting, configudation updates. Tools like crashanalytics solving crash reporting are more valuable for all app developers. Similarly, there are lot of cross cutting concerns we have in the nascent LLM application space.</p><p>Some of the opportunities or cross cutting concerns for LLM application developers(or opportunities for building new start ups).</p><h4><strong>1) Privacy:</strong></h4><p>Lets say you built an agentic workflow to help customer book a flight. In the interactions with user, <strong>How do you ensure that we are not showing User A data to User B? </strong>This gets complicated with valid RBAC controls which allows user to see other users data. E.g. manager checking employee’s time sheets.</p><h4>2) Monitoring/Observability:</h4><p>In LLM world, if it works, user sees a delightful experience. More often, user needs multiple interactions with the system to accomplish their tasks. As application developer <strong>how do I know that in this specific interaction user had a delightful experience(without asking for explicit feedback)?</strong></p><p>In case user had bad experience, how can I trace it without collecting personally identifiable info or the entire chat history?</p><h4>3) Security/filtering:</h4><p>Input validation involves preventing injection attacks. Output filtering involves avoiding harmful/sensitive information. Underlying models and LLM providers(openAI, Azure etc), provide a level of support. However, there are industry or company specific usecases like, in health care industry would like to avoid sharing financial info over chat.</p><h4>4) Model versioning / adaptability:</h4><p>How can we switch the underlying model while keeping my application usecases same? If I change OpenAI model to Claude, how does it change my chat bot experience? A deterministic way of knowing the impact would give a great capability to switch models and reduce costs.</p><h4>5) Compliance:</h4><p>All tools companies build/use ultimately need to adhere to regulations. How do I know that the app I build satisfies say GDPR, are there any easier way to verify the compliance?</p><p>All these would be a wrappers over LLMs, as long as wrappers provide business value and intertwined into niche workflows, they get more valuable. I am sure there will be more successful startups in this space in near future.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=17ca29edc92d" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Validate your trading idea with Chat GPT and Pine script]]></title>
            <link>https://medium.com/@sarath.hotspot/quickly-prototyping-your-trading-idea-with-chat-gpt-and-pine-script-45a2a5a4deb6?source=rss-c828a6bcd49a------2</link>
            <guid isPermaLink="false">https://medium.com/p/45a2a5a4deb6</guid>
            <category><![CDATA[trader-view]]></category>
            <category><![CDATA[algo-trading-strategies]]></category>
            <category><![CDATA[ai]]></category>
            <category><![CDATA[chatgpt]]></category>
            <category><![CDATA[pine-script]]></category>
            <dc:creator><![CDATA[Sarath]]></dc:creator>
            <pubDate>Thu, 28 Nov 2024 22:42:41 GMT</pubDate>
            <atom:updated>2024-11-29T02:45:15.428Z</atom:updated>
            <content:encoded><![CDATA[<p>Do you track trading charts and you are able to find subtle patterns in price movements and wonder whether this pattern works in real trading?</p><p>Over the last couple of years, I dabbled in trading and algo trading. I jotted down lot of trading ideas(E.g. Price tend to increase when there is a strong up tick with big volume) but I could not find a reliable way to quantify whether these ideas work or not.</p><p>I came across PineScript a scripting language provided by TraderView to write indicators and strategies. I have a trading idea, but I am no expert in writing pine script to validate this idea. I thought lets take help from chatGPT.</p><p><strong>My trading hypothesis:</strong> First 5–30 minutes is where we see most volume on a stock, during the day, if there is a 5 minute candle bar with uptick and with more volume than early morning session, price tend to go higher.</p><p><strong>Prompt I gave to ChatGPT:</strong></p><pre>1) Collect first 5 minute candle volume, open, close, high and low. lets call it open candle.<br>2) During the day if we see a price up candle with more volume that first candle volume, we expect price to go up. lets call it trigger candle. <br>3) Place a buy order. take profit order is order price + trigger candle length. stop loss order is order price - trigger candle length. </pre><pre>//@version=5<br>strategy(&quot;Volume-Based Trigger Strategy&quot;, overlay=true, shorttitle=&quot;Volume Trigger&quot;)<br><br>// Parameters<br>var float open_candle_high = na<br>var float open_candle_low = na<br>var float open_candle_volume = na<br>var float open_candle_open = na<br>var float open_candle_close = na<br>var bool open_candle_collected = false<br><br>// Time Settings<br>start_time = timestamp(year, month, dayofmonth, 09, 30) // Start of session<br>end_time = timestamp(year, month, dayofmonth, 09, 35)   // End of first 5 minutes<br><br>// Collect First 5-Minute Candle Data<br>if (time &gt;= start_time and time &lt; end_time)<br>    open_candle_high := high<br>    open_candle_low := low<br>    open_candle_volume := volume<br>    open_candle_open := open<br>    open_candle_close := close<br>    open_candle_collected := true<br><br>// Trigger Candle Conditions<br>trigger_candle = close &gt; open and volume &gt; open_candle_volume and open_candle_collected<br><br>// Trigger Candle Length<br>trigger_candle_length = high - low<br><br>// Place Buy Order with Take Profit and Stop Loss<br>if (trigger_candle)<br>    entry_price = close<br>    take_profit = entry_price + trigger_candle_length<br>    stop_loss = entry_price - trigger_candle_length<br>    strategy.entry(&quot;Buy&quot;, strategy.long, stop=entry_price)<br>    strategy.exit(&quot;Take Profit&quot;, &quot;Buy&quot;, limit=take_profit, stop=stop_loss)<br><br>// Reset for Next Day (if needed)<br>if (time &gt;= timestamp(year, month, dayofmonth, 15, 55))  // End of trading session<br>    open_candle_high := na<br>    open_candle_low := na<br>    open_candle_volume := na<br>    open_candle_open := na<br>    open_candle_close := na<br>    open_candle_collected := false</pre><p>I asked to add markers on trigger bars for analysis.</p><pre>Could add arrows when we are collecting first candle information. <br>And for trigger candle. </pre><p>No so surprisingly… this script was executed perfectly in traderview.</p><p>I ran this script on MSTR for 2024–08–26 to 2024–11–27.</p><p>It generated 59% success trades.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*UVuAM9llAWtIDQthCIU5bA.png" /></figure><p><strong>What I learned from this experiment:</strong></p><ol><li>I could quickly proto type my trading ideas in couple of minutes. It gives people a powerful tool to iterate and find their own edge in market.</li><li>ChatGPT has grown a lot in last couple of years: LLMs used to generate syntactically correct code but it requires some manual tweaks to make it right. But in recent days, LLMs are generating functionally working code.</li></ol><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=45a2a5a4deb6" width="1" height="1" alt="">]]></content:encoded>
        </item>
    </channel>
</rss>