Full-Stack Developer AI Integration London, UK

I build
AI-powered
SaaS apps that actually ship to production.

Production-grade RAG systems, embeddable chatbots, and workflow automation for founders, ops teams, and SaaS companies that need more than a prototype.

Explore work → Start a project
4+ Products shipped
40% Support reduction
24h Typical response
Specialism
AI + MERN
Based in
London, UK
Status
Open for work
01 About

I'm a full-stack developer based in London, building production AI applications for founders and small teams. Over the past two years I've shipped four AI tools running live: a multi-tenant document Q&A platform, an embeddable customer-support widget with analytics, an email-to-Slack automation system, and an EdTech study app.

My focus is the gap between working in a notebook and working under real users. Authentication, billing, multi-tenancy, deployment, and observability are part of the deliverable, not afterthoughts.

I work across UK and US timezones, take on a small number of projects at a time, and communicate in plain English about scope and trade-offs.

Frontend

  • React
  • Next.js
  • TypeScript
  • Tailwind

Backend

  • Node.js
  • Express
  • FastAPI
  • MongoDB
  • PostgreSQL

AI / RAG

  • OpenAI
  • Claude
  • LangChain
  • ChromaDB
  • Vector search

Infra

  • AWS
  • Vercel
  • Railway
  • Docker
  • Stripe
  • OAuth
02 Client Results

Proven delivery for teams that needed speed, quality, and real business outcomes.

Byte Bricks — Engagement

Full Stack Developer — Byte Bricks

January 2025 - August 2025

  • Shipped 3 production MERN applications serving live users end-to-end.
  • Containerised backend services with Docker on AWS Fargate, cutting environment-related bugs by 40%.
  • Built CI/CD pipelines from scratch, reducing deployment time by 30% and increasing feature velocity.
  • Optimised 15+ REST endpoints; average response time improved 25% through query and cache work.

Previously: MERN engineering at HATTRICK Solutions (2024).

03 Selected Work

Four production tools, each solving a real workflow problem.

Every project below is live and demoable, with real product screenshots from shipped builds. Click through to the live apps for the full workflow and interaction quality.

DashBot analytics dashboard screenshot
Project / 01

DashBot

Embeddable AI customer-support chatbot with a real analytics dashboard. Tracks unanswered questions, routing breakdown, and per-bot usage.

Problem
Small SaaS teams burn hours on repetitive support tickets and have no visibility into what users actually ask.
Approach
RAG over uploaded docs, one-line embed script, full dashboard for analytics and unanswered intents.
Outcome
Reduces routine support volume by ~40% based on internal testing; usable in any website with a single script tag.
Stack
ReactViteTypeScriptFastAPIMongoDBQdrantGroqDockerRailwayTailwind CSSshadcn/ui
View live →
RAG Workspace application screenshot
Project / 02

RAG Workspace

Multi-tenant internal knowledge base for companies. Admins upload documents, users chat with them, and each workspace stays fully isolated.

Problem
HR, ops, and compliance teams need grounded answers from internal docs — not generic AI responses.
Approach
FastAPI backend with ChromaDB vector store, React frontend, role-based admin console, JWT auth.
Outcome
Per-workspace document isolation, per-user chat history, admin console for managing teams and upload permissions.
Stack
FastAPIReactChromaDBMongoDB
View live →
InboxAI automation dashboard screenshot
Project / 03

InboxAI

Email automation that connects Gmail to Slack, Notion, and ClickUp. AI classifies urgency and routes work without manual triage.

Problem
Ops teams and founders drown in inbox triage; important client emails get lost, low-priority noise gets too much attention.
Approach
Gmail OAuth, AI classification of each new email, conditional routing to Slack alerts, Notion tasks, or ClickUp tickets.
Outcome
Preview mode for safe testing, configurable auto-run intervals, dashboard for workflow monitoring and health.
Stack
Next.jsReactTypeScriptMongoDBJWTGoogle OAuthSlack APINotion APIClickUp APITailwind CSSRailway
View live →
Lecture Study learning interface screenshot
Project / 04

Lecture Study

EdTech platform that turns any YouTube lecture into a structured study session with summaries, quizzes, and a per-lecture Q&A chat.

Problem
Students and self-learners want study materials from YouTube lectures, not just raw transcripts.
Approach
Pulls public captions, generates AI study guides and MCQs, layers a knowledge-check chat scoped to each lecture.
Outcome
Per-lecture saved sessions, study guides and key ideas on demand, quizzes with explanations, light/dark UI.
Stack
ReactNode.jsOpenAIYouTube APIMongoDB
View live →
04 Services

What I build for clients.

05 Contact

Have a project in mind?

imjunaidnoor@gmail.com →
Fiverr Email View projects