Projects & reference implementations

Demonstrated builds and technical implementations.

Below is an honest view of our work. Each entry clearly identifies whether it is a completed website build, an internal infrastructure blueprint, or an experimental prototype.

We describe the practical requirements, technologies used, and engineering decisions behind each project without exaggerating outcomes or publishing unverified statistics.

Aama Pathshala — bilingual preschool website

Completed web build
Education / EN + NP
Aama Pathshala preschool website interface showing facilities and contact points

A responsive website built for Aama Pathshala, a preschool in Nepal, providing parents with clear information on facilities, classroom amenities, and admissions contact.

Project requirements

Parents looking for early childhood schooling need direct, practical information: classroom environment details, air conditioning, CCTV safety, transportation bus routes, and straightforward contact numbers. The site needed to serve families comfortably in both Nepali and English, with layouts tailored for mobile smartphone screens.

What was built

A lightweight static website featuring a prominent amenities grid, an admission enquiry section, direct telephone and WhatsApp links, and an in-page toggle allowing users to switch between English and Nepali text without page reloads.

Technical implementation decisions

  • Static implementation: Built with semantic HTML5, modern CSS, and plain JavaScript. Avoiding heavy JavaScript application frameworks kept total page weight small for mobile visitors.
  • Client-side language toggle: Implemented a clean DOM data-attribute switcher that toggles text between English and Nepali and saves the preference in browser storage.
  • Touch-friendly actions: Configured telephone, WhatsApp, and location map links as clear touch targets for mobile visitors.
Project type
Educational website build
Stack
HTML5, CSS3, vanilla JavaScript
Key features
Bilingual toggle, mobile layout
Status
Completed build

Linux server baseline & automated deployment pipeline

Infrastructure blueprint
Linux / Ubuntu
Reference CaddyfileBlueprint / not a live status report
yugya.com, www.yugya.com {
    root * /var/www/yugya
    file_server
    encode zstd gzip
    header {
        Strict-Transport-Security "max-age=31536000; includeSubDomains; preload"
        X-Content-Type-Options "nosniff"
        X-Frame-Options "DENY"
        Referrer-Policy "strict-origin-when-cross-origin"
    }
}

A standardized reference configuration for deploying web applications on Ubuntu Linux virtual private servers, featuring firewall rules, reverse proxying, and automated Git-based deployment.

The operational goal

When deploying web projects to cloud VPS providers (such as DigitalOcean or Linode), default server setups often leave root password authentication enabled and firewall rules unconfigured. Manual file uploads via FTP also create risk of incomplete transfers. This blueprint establishes a clean, repeatable deployment baseline.

What was configured

  • Documented server setup on Ubuntu LTS.
  • Enforced SSH public-key authentication with root password login disabled.
  • Uncomplicated Firewall (UFW) allowing only HTTP (80), HTTPS (443), and designated SSH traffic.
  • Caddy reverse proxy providing automated TLS certificate issuance and renewal via Let's Encrypt.
  • Automated GitHub Actions deployment pipeline that syncs repository updates cleanly on merge.

Technical implementation decisions

  • Caddy reverse proxy: Selected Caddy for its reliable automated Let's Encrypt certificate lifecycle management and clean, concise configuration format.
  • Standard security headers: Configured HSTS, nosniff, and frame restrictions in the proxy configuration to defend against MIME-type sniffing and framing.
  • Git-based publishing: Configured automated deployment via GitHub Actions, replacing manual file transfers with atomic, repeatable updates.

Scope note

This is a reference baseline, not a claim of guaranteed uptime or a managed hosting product. Server accounts, billing, and access remain under the client's ownership.

System type
Server baseline & blueprint
Operating system
Ubuntu Linux LTS
Proxy & TLS
Caddy / Nginx, Let's Encrypt
Status
Active reference setup

AayushAI Core — terminal assistant with SQLite memory

Personal prototype
Python / SQLite
Prototype outputLocal CLI / illustrative excerpt
$ aayushai --status
[SQLite]  Database: ~/.aayushai/memory.db (connected)
[Records] Structured notes and task reminders loaded.
[Execution] Local CLI process running on host machine.

A personal research prototype exploring persistent, structured memory for command-line developer utilities using a local SQLite database.

The exploration

Most interactive command-line helpers and chat interfaces lose context as soon as the terminal window is closed. This experiment investigated whether using standard relational database tables on local disk can provide a clean, queryable persistence layer for notes, reminders, and ongoing project tasks without cloud dependencies.

What was built

A standalone Python CLI application backed by a structured SQLite database file. Users can add reminders, save developer journal notes, and search past context directly from the terminal.

Technical implementation decisions

  • Local database schema: Structured data tables in SQLite allow indexing and standard SQL queries across sessions rather than relying on unindexed flat text files.
  • Zero cloud requirement: The database file lives locally on the user's computer, keeping notes private and fully inspectable.
  • Standard library focus: Built using Python standard libraries (sqlite3, argparse) to ensure ease of execution on Linux and macOS environments.
Category
Personal prototype / tool
Languages
Python 3, SQL
Storage
Local SQLite3 database
Status
Functional experiment
Portfolio transparency

Clear representation of our work.

This page presents projects, infrastructure reference setups, and personal prototypes that have been built and verified by our team.

We do not publish unverified client logos, unconfirmed partnerships, or invented outcome metrics. Where work involves private client systems or confidential requirements, we discuss technical architecture and approaches directly during scoping conversations.

If you would like to review code repositories, technical configurations, or discuss relevance to your project, we welcome a direct conversation.

Have a project requirement to discuss?

Tell us about your website, server setup, or technical requirement. We will review your goals and provide an honest assessment of scope and architecture.

Direct email
hello@yugya.com
Response aim
1–2 business days
Location
Nepal · UTC +5:45