<devvrania/>

kairo ~ compiling Rania's portfolio▍

Kairo, Rania's coding companion, at his desk in a night room with a cherry tree outside the window

Hello! I'm

RaniaNasir

I build software that feels thoughtful.

See my workLinkedIn(opens in a new tab)GitHub(opens in a new tab)

01 · about

About me

I’m a software engineer with two years of production experience building scalable full-stack and backend applications with Node.js, NestJS, React, TypeScript and AWS.

Most of my work is REST APIs, microservices and data-heavy systems, with a focus on performance and maintainability. Lately that includes AI-powered features built on LLMs and retrieval-augmented generation.

I care about clean architecture, maintainable code, and the little details that make software better.

The character at the top is Kairo, my 3D coding companion. Move your cursor over his room and he'll follow it; click him and he'll say hi.

faster API responses, through Redis caching and query work
30%faster API responses, through Redis caching and query work
faster MongoDB and PostgreSQL queries, through indexing
40%faster MongoDB and PostgreSQL queries, through indexing
less code complexity, refactoring legacy Node.js services
20%less code complexity, refactoring legacy Node.js services

02 · experience

Where I've worked

  1. Software Engineer · Devsinc

    July 2024 – Present · Pakistan
    • Built and maintained production full-stack applications with Node.js, NestJS, React, Next.js and AWS, across multiple production deployments
    • Designed scalable backend services and microservice architectures, integrating third-party APIs for business-critical workflows
    • Built an AI-powered RAG system with Firebase Genkit and LLMs for querying 300+ employee records and organization-specific HR data
    • Cut API response times by up to 30% with Redis caching, query optimization and better data access
    • Improved MongoDB and PostgreSQL query performance by up to 40% through indexing and query optimization
    • Reduced code complexity by 20% by refactoring legacy Node.js services
    • Worked with QA and DevOps to troubleshoot production issues and keep downtime low
    Node.jsNestJSReactNext.jsAWSRedisMongoDBPostgreSQL

03 · projects

Things I've built

Client work

bizscout.com

Bizscout

bizscout.com ↗ (opens in a new tab)

A business acquisition marketplace connecting buyers, sellers and brokers through centralized listings and broker management workflows.

Next.jsNestJSMongoDBCRM integrationsSupabase+5
bizscout.com

Client work

Bizscout

bizscout.com ↗ (opens in a new tab)

A business acquisition marketplace connecting buyers, sellers and brokers through centralized listings and broker management workflows.

What I did

  • Built and maintained 10+ automated Firecrawl scrapers processing 25,000+ business listings a day
  • Designed the daily pipeline that sanitizes, normalizes, standardizes and scores scraped data, with retries, failure handling and rate-limit management
  • Built the supply team’s admin dashboard to review, approve and edit listings, manage leads and monitor scrapers
  • Integrated brokerage partners to ingest for-sale listings, with automated broker notifications and lead tracking when buyers show interest
  • Built CRM integrations for the broker and lead workflows
  • Mapbox geocoding and Resend for personalized automated email workflows

Stack

Next.jsNestJSMongoDBCRM integrationsSupabaseAWSDockerRailwayVercelGitHub Actions
heloufirm.com

Helou Firm – MyFileHR

heloufirm.com ↗ (opens in a new tab)

A full-stack HR platform for employee onboarding, leave management, feedback and organization-specific workflows, with a RAG-powered HR assistant.

RAGFirebase GenkitGeminiReactNode.js+3
heloufirm.com

Client work

Helou Firm – MyFileHR

heloufirm.com ↗ (opens in a new tab)

A full-stack HR platform for employee onboarding, leave management, feedback and organization-specific workflows, with a RAG-powered HR assistant.

What I did

  • Built a multi-organization RAG chatbot with Firebase Genkit, Gemini and Firestore, serving 5+ organizations and 300+ employee records
  • AI query routing that works out what is being asked and retrieves from that organization’s HR data and policies
  • Role-based access control for admins, managers and employees, so AI answers only use data the asker is allowed to see
  • Personalized questions about leave balances, feedback, employee details and company policy

Stack

RAGFirebase GenkitGeminiReactNode.jsReduxFirebaseFirestore
rewaa.com

Rewaa

rewaa.com ↗ (opens in a new tab)

A retail management platform, where I worked as a full-stack engineer on frontend features, backend services and API integrations.

AngularNode.jsAWSPostgreSQLDocker
rewaa.com

Client work

Rewaa

rewaa.com ↗ (opens in a new tab)

A retail management platform, where I worked as a full-stack engineer on frontend features, backend services and API integrations.

What I did

  • Built the onboarding feature end to end, from the frontend flows to the APIs behind them
  • Integrated and maintained APIs between the frontend, backend and external systems

Stack

AngularNode.jsAWSPostgreSQLDocker

Personal projects

books.devvrania.com

Book Intelligence

books.devvrania.com ↗ (opens in a new tab)Code ↗ (opens in a new tab)

An open, AI-powered library. Upload a PDF and ask it anything: every answer comes only from the book, with citations to the exact chapter and page.

Next.jsTypeScriptRAGpgvectorPostgreSQL+3
books.devvrania.com

Personal project

Book Intelligence

books.devvrania.com ↗ (opens in a new tab)Code ↗ (opens in a new tab)

An open, AI-powered library. Upload a PDF and ask it anything: every answer comes only from the book, with citations to the exact chapter and page.

What I did

  • Hybrid RAG: pgvector similarity plus Postgres full-text search, and citations can only point at passages that were actually retrieved
  • Background ingestion pipeline: PDF extraction, cleaning, chapter detection, structure-aware chunking and embeddings, resumable after failures
  • Runs on local Ollama models or any OpenAI-compatible API; the live demo uses Gemini’s free tier and waits out its per-minute limits
  • Deployed for free on Render, Neon (Postgres + pgvector) and Supabase Storage, with no accounts and no personal data stored

Stack

Next.jsTypeScriptRAGpgvectorPostgreSQLDrizzleGeminiOllama
kairo.devvrania.com

Kairo

kairo.devvrania.com ↗ (opens in a new tab)Code ↗ (opens in a new tab)

My 3D coding companion: a VRM character at a night desk who types, follows your cursor, takes stretch breaks and says hello when you click him. He’s the one at the top of this page.

three.jsthree-vrmJavaScriptWebGL
kairo.devvrania.com

Personal project

Kairo

kairo.devvrania.com ↗ (opens in a new tab)Code ↗ (opens in a new tab)

My 3D coding companion: a VRM character at a night desk who types, follows your cursor, takes stretch breaks and says hello when you click him. He’s the one at the top of this page.

What I did

  • Procedural seated pose built from the model’s own bones, with breathing, typing and gaze layered on top
  • Mood, attention and idle-break behaviours driven by what the visitor is doing
  • The whole room is generated in code: canvas-painted window, IDE screen and keyboard backlight

Stack

three.jsthree-vrmJavaScriptWebGL

04 · skills

Tech I work with

Languages

TypeScriptJavaScriptSQL

Frontend

ReactNext.jsAngularReduxTailwind CSSMaterial UIthree.js

Backend & APIs

Node.jsNestJSExpressREST APIsWebSocketsJWT authMongooseBullMQRedisDrizzle ORM

Databases

MongoDBPostgreSQLMySQLSupabasepgvector

AI & Integrations

CRM integrationsLLM integrationsRAGOllamaFirebase GenkitGeminiFirecrawlMapboxStripeSendGridResend

Cloud & DevOps

AWSDockerGitHub ActionsCI/CDVercelRailwayRenderFirebase

Architecture

MicroservicesEvent-drivenMVCSOLID

Tools

GitGitHubGitLabJiraLinearClaude Code

05 · contact

Let's build something

I'm always happy to talk about full-stack work, AI-powered products, or a project you're planning. Email or LinkedIn both reach me.

devv.rania@gmail.comLinkedIn(opens in a new tab)GitHub(opens in a new tab)