SShortSingh.
TopTechnologyAI & MLStartupsProgrammingScienceBusinessFinanceCrypto & Web3World
Back to feed
Programming

CS Student Deploys React Portfolio on AWS S3 and CloudFront, Learns From Errors

0
·1 views

A computer science student documented their first hands-on experience building and deploying a personal portfolio using React, AWS S3, and CloudFront as part of a self-directed cloud workshop. The project began as a simple HTML file before being converted into a React and Vite application with structured components for skills, projects, and education. During deployment, the student encountered an S3 AccessDenied error caused by missing public-access settings and an incorrect bucket policy referencing the wrong bucket name. A further CloudFront misconfiguration added to the debugging process, ultimately requiring the student to trace each error back to its root cause. The experience highlighted the importance of understanding cloud infrastructure configuration rather than copying settings without context.

Read the full story at DEV Community

This is an AI-generated summary. ShortSingh links to the original source for the complete article.

Discussion (0)

Log in to join the discussion and vote.

Log in

Related stories

0
ProgrammingDEV Community ·

Developer Builds Custom MCP Server to Let Managers Query 20 Years of Azure SQL Data via Claude

A developer built a custom Model Context Protocol (MCP) server to allow company managers across departments like HR, Sales, and Production to query two decades of business data using Claude, without any SQL knowledge or BI tools. The system routes requests from Claude through a self-hosted Node.js server on the client's own VPS, then to an Azure SQL database via a read-only SQL user, keeping all data off third-party infrastructure. A table whitelist built into the MCP layer ensures Claude can only access approved tables, preventing exposure of legacy, staging, or sensitive data. The server exposes three core tools to Claude: listing available tables, describing table schemas, and running SELECT-only queries capped at 500 rows. Key design choices included self-hosting for data privacy, enforced read-only database access to prevent destructive queries, and a shared secret key for authenticating manager accounts.

0 comments Read more at DEV Community
0
ProgrammingDEV Community ·

TraceEagle Tool Lets Developers View, Decode, and Preview Captured Network Traffic

TraceEagle is a network traffic analysis tool that allows developers to inspect captured HTTP requests and responses in multiple viewing modes, including Structured, Pretty Text, Hex, and Auto Detect. The tool automatically decompresses data encoded with gzip, brotli, deflate, or zstd, and pretty-prints formats such as JSON, XML, forms, and protobuf without requiring additional configuration. Users can preview media files including images, video, and audio directly within the interface instead of saving them separately. Each request detail panel displays the request and response independently, allowing different view modes to be applied to each side simultaneously. For binary or proprietary protocols that standard views cannot parse, the tool also supports custom protocol decoding.

0 comments Read more at DEV Community
0
ProgrammingDEV Community ·

LLM Wiki Replaces Traditional RAG With Cached, Traceable Wiki Page Generation

LLM Wiki is a cross-platform desktop application that takes a two-phase approach to document ingestion, first analyzing sources to extract entities and relationships, then generating persistent wiki pages with full source references. Unlike traditional RAG systems that re-embed documents on every query, LLM Wiki caches intermediate analysis results so incremental updates cost significantly less — roughly $0.02 per update versus $0.50 for a full re-analysis. Every generated wiki page links back to its original source file, whether a PDF, image, or web clip, and deletions or changes automatically propagate through the knowledge graph. The ingest queue includes crash recovery, resuming from the last checkpoint if the application is interrupted mid-process. The project has garnered over 19,500 stars on GitHub and was recently trending as the tenth most popular TypeScript repository on the platform.

0 comments Read more at DEV Community
0
ProgrammingDEV Community ·

How to Properly Test an AI Spreadsheet Assistant Before Trusting Its Numbers

A practical guide urges users to test spreadsheet AI tools against questions with already-known answers before relying on them for real business analysis. The recommended method involves creating a sanitized copy of an actual workbook — retaining messy real-world details like blank rows, duplicate IDs, and ambiguous labels — then running at least ten pre-verified questions. Testers should evaluate not just whether the final number looks correct, but whether the AI used the right range, filters, and definitions to arrive at it. The guide also compares tools across ecosystems, including Excel Copilot, Google Sheets AI, WPS AI, Teable, and AI SQL translators, noting that the best choice typically depends on where the data already lives. A core principle applies across all tools: the AI proposes a formula or query, but the human remains responsible for verifying the result.

0 comments Read more at DEV Community
CS Student Deploys React Portfolio on AWS S3 and CloudFront, Learns From Errors · ShortSingh