Muhammed Suhail EA | Full Stack & AI Engineer

01 — About

Reliable systems, where backend depth meets applied AI.

I'm a full stack engineer with 6+ years building production software for enterprise clients — HR & payroll platforms, offline-first attendance PWAs, and conversational AI. My work spans the whole stack, but I'm most at home in the backend: microservices, message queues, data modelling, and the architecture that keeps things fast at scale.

Lately I focus on applied AI — LLM agents with LangGraph, semantic search over Qdrant, and RAG pipelines that put real reasoning behind a chat box. I care about clean APIs, observable systems, and shipping things people actually use.

Currently at HITEK AI in Dubai, building enterprise AI for Microsoft Teams.

suhail@dev: ~/whoami

// whoami

const engineer = {

role: 'Full Stack & AI Engineer',

at: 'HITEK AI · Dubai',

focus: ['Backend architecture', 'LLM agents', 'RAG'],

stack: ['NestJS', 'Next.js', 'Python', 'FastAPI', 'LangGraph'],

}

6+ years·99.9% uptime SLA

02 — Stack

The toolkit.

{ }

Frontend

TypeScriptReactNext.jsVue 2/3NuxtSvelteKitTailwindPinia

</>

Backend

Node.jsNestJSHono.jsDenoPythonFastAPIGraphQL / ApolloPrismaTypeORMRabbitMQBullMQ

AI & LLM

LangGraphLangChainVercel AI SDKMCPGPT-4oClaudeRAGAI AgentsQdrantFunction Calling

Data & Storage

PostgreSQLMySQLMS SQLMongoDBRedisSupabaseQdrantIndexedDB

☁

DevOps & Cloud

DockerKubernetesGitHub ActionsCI/CDAWSAzureVercelNginx

▤

Architecture

MicroservicesEvent-DrivenTurborepoPWAOAuth 2.0Azure MSALSystem Design

03 — Experience

Where I've worked.

Jul 2025 — PresentDubai, UAE

Full Stack & AI Engineer

HITEK AI

Delivered the Maaden Teams AI chatbot for enterprise service management.
Implemented semantic search with the Qdrant vector database.
Optimized performance using Redis caching and BullMQ job queues.
Built end-to-end approval workflows inside Microsoft Teams.

Vue.jsNestJSLangGraphQdrantTeams AI SDK

Oct 2021 — Jun 2025Dubai, UAE

Full Stack Developer

RADIXHR

Architected the core payroll calculation engine with overtime and variable-pay logic.
Implemented GraphQL APIs serving 10,000+ users.
Reduced query response time by 60% through optimization.
Led migration to Kubernetes for high-availability deployment.

Vue 3NestJSGraphQLApolloTypeORMMySQL

Apr 2020 — Sep 2021India

Software Engineer

Pulse Star Technical Solutions

Launched Radix-TAS with offline-first PWA architecture and IndexedDB.
Integrated enterprise security with Azure MSAL and OAuth 2.0.
Delivered mobile-responsive interfaces with Vue.js and Element UI.
Set up a containerized deployment pipeline with Docker.

Vue 2ReactNode.jsExpressSQL Server

04 — Selected work

Five systems, shipped to production.

AI · FEATURED01

Maaden Teams AI Chatbot

Conversational AI inside Microsoft Teams — service requests, approvals and CAFM, reasoned with LangGraph and grounded by Qdrant semantic search.

Multi-stepLangGraph reasoning

SemanticQdrant retrieval

Vue.jsNestJSLangGraphQdrantBot Framework

Maaden Assistant · ● online

Raise a service request for AC repair in Block B

ADAPTIVE CARD

Request #SR-1042
AC repair · Block B · High

ApproveReassign

ENTERPRISE02

Radix HR

Enterprise payroll engine with overtime and variable-pay calculation, built on microservices for scale and reliability.

10k+active users

99.9%uptime

<200msAPI response

Vue 3NestJSGraphQLApolloTypeORM

Payroll · March 2026 · Processed

Gross pay$184,250

Overtime$12,840

Deductions−$9,120

Net payout$187,970

AI03

Flowtrail AI

Agent-builder platform with natural-language-to-SQL and RAG, orchestrating 6+ LLM providers behind one interface.

6+LLM providers

RAGgrounded answers

Next.jsTypeScriptLangChainLangGraph

Agent Builder

OpenAIClaudeGeminiGrokMistral

// natural language → SQL

"top 5 regions by Q1 revenue"

SELECT region, SUM(revenue) FROM sales WHERE q = 1 GROUP BY region LIMIT 5;

ENTERPRISE04

Radix-TAS

Offline-first time-attendance PWA with IndexedDB persistence, Azure MSAL auth, and real-time sync when the network returns.

Offlinefirst architecture

MSALenterprise auth

Vue 2ReactNode.jsExpressSequelize

Live Attendance · OFFLINE · SYNCED

AK Arun KumarInside · 5:42

SN Sneha NairInside · 5:30

JT Joseph TCheck in

TOOL05

SQL Copilot

Natural-language SQL assistant powered by GPT-4o-mini via the Vercel AI SDK, with real-time streaming responses.

70%faster queries

Streamingresponses

Next.js 15TypeScriptTailwindOpenAIVercel AI SDK

SQL Copilot

Top 10 customers by revenue this year

SELECT customer, SUM(amount) AS rev FROM orders WHERE year = 2026 GROUP BY customer ORDER BY rev DESC LIMIT 10;

MuhammedSuhail EA

Reliable systems, where backend depth meets applied AI.

The toolkit.

Where I've worked.

Five systems, shipped to production.

Let's build something that ships.