<?xml version="1.0" encoding="utf-8" standalone="yes"?><rss version="2.0" xmlns:atom="http://www.w3.org/2005/Atom"><channel><title>Software &amp; AI | Shivam Vashishtha</title><link>https://www.shivamvashishtha.com/tag/software-ai/</link><atom:link href="https://www.shivamvashishtha.com/tag/software-ai/index.xml" rel="self" type="application/rss+xml"/><description>Software &amp; AI</description><generator>Wowchemy (https://wowchemy.com)</generator><language>en-us</language><lastBuildDate>Tue, 01 Jul 2025 00:00:00 +0000</lastBuildDate><image><url>https://www.shivamvashishtha.com/media/icon_hu016cb3b331e9522dbd920334fc7fb877_32429_512x512_fill_lanczos_center_3.png</url><title>Software &amp; AI</title><link>https://www.shivamvashishtha.com/tag/software-ai/</link></image><item><title>Scendro — Financial Analytics Prototype</title><link>https://www.shivamvashishtha.com/project/stockscope/</link><pubDate>Tue, 01 Jul 2025 00:00:00 +0000</pubDate><guid>https://www.shivamvashishtha.com/project/stockscope/</guid><description>&lt;p>&lt;strong>Context:&lt;/strong> Independent software project&lt;br>
&lt;strong>Period:&lt;/strong> Prototype and iterative development&lt;br>
&lt;strong>Focus:&lt;/strong> Software &amp;amp; AI&lt;br>
&lt;strong>Tools / methods:&lt;/strong> React / TypeScript · FastAPI · Python · market-data analysis&lt;/p>
&lt;h2 id="engineering-question">Engineering question&lt;/h2>
&lt;p>A financial research interface benefits from separating data acquisition, transformation, model logic, and display so that analysis assumptions remain understandable.&lt;/p>
&lt;h2 id="my-work">My work&lt;/h2>
&lt;ul>
&lt;li>Worked on a web application architecture for market-data research and visualization.&lt;/li>
&lt;li>Built or iterated on interface and backend workflows, including API-driven analytics.&lt;/li>
&lt;li>Explored portfolio views and experimental machine-learning-assisted analysis.&lt;/li>
&lt;/ul>
&lt;h2 id="technical-approach">Technical approach&lt;/h2>
&lt;p>A general architecture consists of a frontend for input and exploration, backend endpoints for validated requests, data loaders with timestamp-aware normalization, and analytical components separate from presentation. Time-series modeling must prevent look-ahead bias and distinguish research experiments from deployable forecasts.&lt;/p>
&lt;h2 id="engineering-considerations">Engineering considerations&lt;/h2>
&lt;p>Market forecasting is vulnerable to nonstationarity, overfitting, and leakage. Backtesting and prediction demonstrations should include transaction assumptions, appropriate holdout periods, and uncertainty, not only visually compelling charts.&lt;/p>
&lt;h2 id="evidence-and-maturity">Evidence and maturity&lt;/h2>
&lt;p>Software prototype and experimental analysis; not a claim of validated alpha, automated trading, or financial advice.&lt;/p>
&lt;h2 id="public-scope">Public scope&lt;/h2>
&lt;p>No private API credentials, proprietary datasets, or unverified return claims are shared.&lt;/p></description></item></channel></rss>