m
mirazxdev

Abdullah Miraz

@mirazxdev

AI RAG and LLM Integrations with Nextjs TypeScript

Bangladesh
English, Bengali, Hindi, Urdu
About me
Software engineer for AI-native apps and backends. Next.js/TypeScript with real AI — RAG over your docs, LLM integrations, APIs — not bolted-on chatbots. EMR at ATI Limited · CMS + automation at KumoDevs · freelance full-stack. MSc SE (STANKIN) · BSc CSE (IUBAT). Work: MindStore AI (RAG), Medibuy, RxPro, Pulse Analytics, Autism Insight (ML). I help with: • RAG / chat-with-your-docs (citations, streaming, BYOK) • OpenAI / Claude / Gemini / DeepSeek into existing apps • Next.js/React, dashboards, PostgreSQL, APIs & automation Message goal + stack before order — I will confirm scope/timeline.... Read more

Skills

m
mirazxdev
Abdullah Miraz
Offline • 
Average response time: 1 hour

See my services

AI Integrations
I will integrate chatgpt openai and llm api into your web app or website
From$90 / projectMore details
AI Integrations
I will build a custom rag ai chatbot trained on your documents and data
From$90 / projectMore details

Portfolio

Work experience

Software Engineer

KumoDevs

Jan 2025 - Present • 1 yr 8 mos

Re-architected Next.js + MDX CMS improving rendering performance and content scalability (25-30% faster page loads). Built automation and scraping systems reducing manual operational workload by 50%. Designed scalable PostgreSQL/Supabase schemas optimized for high-content systems.

ATI_Limited

Software Engineer

ATI Limited

Jun 2024 - Dec 2024 • 6 mos

Improved API latency by 40% in production EMR system via query optimization and reduced DB roundtrips. Introduced indexed database queries for high-frequency datasets improving lookup performance (35-50%). Built dashboard-heavy enterprise systems with optimized pagination, caching, and role-based workflows.

Software Engineer

Code Elevate Tech • Full-time

May 2024 - Jun 2024 • 1 mo

Optimized Next.js applications improving UI responsiveness and reducing re-renders (15-20%). Refactored API integration layer improving separation between frontend and backend logic.