United Arab Emirates, Data Engineering & AI systems

Muhammed
Shazin C. M

Data Science & AI Engineering at IIT Guwahati. Founder and sole developer of Kognition, UAE's first AI-native business intelligence layer invented to live on top of existing business systems for UAE SMEs delivering daily revenue insights directly onto WhatsApp in place of a dedicated data analyst.

Muhammed Shazin C. M
K
KognitionFounder · Developer
IIT GuwahatiData Science & AI
Progressing

BS (Hons) DS & AI, IITG

Building

Kognition

Based

Abu Dhabi, UAE

Focus

Applied AI · BI · Data eng.

About

I build data & AI systems end to end, then ship them.

I'm a Data Science & AI Engineering undergraduate at IIT Guwahati, living and working in Abu Dhabi. Most of what I know I learned by building: data pipelines, BI dashboards, LLM applications, and the infrastructure that keeps them running.

Right now that means Kognition. Built solo, 244 days, from the first architecture decision through ingestion, analytics, delivery, the website, and the admin tooling, while studying full time.

Before that, alongside full-time study, I worked in data roles across healthcare, transport logistics, and national-scale events in the UAE. That work taught me the unglamorous part of the job: cleaning, validating, and refusing to report a number I can't trust.

NameMuhammed Shazin C. M
DegreeBS (Hons) Data Science & AI Engineering
UniversityIIT Guwahati
CompanyKognition
LocationAbu Dhabi, UAE
Featured build

Kognition

Business intelligence for UAE SMEs that doesn't need anyone to open a dashboard.

Intelligence that works while you sleep.

Founder & sole developer · with co-founder Amal Moidu Mohamed (operations & commercial)

● Built & deployed · first clients in onboarding

Small business owners in the UAE usually check the till at closing and guess. Kognition reads the daily sales export they already have, works out what changed, and sends the owner a short brief on WhatsApp every morning: revenue, orders, margin, best-seller, and an alert if something looks off. They don't change their POS or install anything.

IngestDaily export from the client's own Google Drive, with flexible per-industry column matching
Sanity-checkHolds the brief when data is incomplete or corrupt, so it never sends a wrong number
KPIsRevenue, orders, margin, best-sellers computed per client
DetectStatistical deviation against each client's own same-weekday rolling baseline
ForecastStatistical projection with day-of-week, UAE holiday & Ramadan seasonality
DeliverClaude writes the plain-language brief, sent via WhatsApp at 07:00 UAE

Stack

  • Python / Flask on Railway (CI deploy from GitHub)
  • Supabase: auth & client portal
  • Twilio WhatsApp · Google Drive API
  • Anthropic Claude API for language generation
  • Netlify: bilingual EN/AR site, RTL, dark/light

Engineering decisions

  • Multi-tenant: one deployment, onboarding a client adds a registry record, not code
  • Client data stays in the client's Drive, with revocable read-only access
  • Admin API secured with a server-side key and constant-time comparison
  • Sanity layer verified in production by holding a zero-revenue brief

Built for

  • F&B and cafés
  • Private clinics
  • Retail
  • Real estate agencies
  • Contracting & transport

An applied-AI system, not a proprietary trained model. The value is in the assembled pipeline, UAE-specific calibration, and the delivery channel.

Experience

Where I've worked

Data roles across events, logistics, and healthcare in Abu Dhabi.

Oct 2025 — Mar 2026
Abu Dhabi · Contract

Data Management Specialist

NG9 Holding (parent company of Nirvana) · Sheikh Zayed Festival

  • Part of a six-person team running data operations for the Sheikh Zayed Festival, one of Abu Dhabi's major cultural events.
  • Built data workflows covering client documentation, system integrations, and automated report generation.
  • Produced reporting that let stakeholders make operational decisions while the festival was running.
Sep 2024 — Feb 2025
Bangalore, India · Internship

Data Engineering Intern · GenAI · ML Pipelines · LLM Architecture

Shiksha Vertex (Skill partner: Microsoft)

  • Generative AI: worked with large language models and prompt-driven workflows as part of the internship's project work.
  • ML pipelines: worked across the stages of a machine-learning pipeline, from data preparation through model training and evaluation.
  • LLM architecture: studied how LLM-based applications are structured, from model APIs to retrieval and orchestration layers.
Mar 2022 — Oct 2022
Abu Dhabi · Part-time

Clinical Data Associate

NMC Health Care

  • Processed and validated patient records against data-quality checks.
  • Applied error-detection and validation protocols to resolve discrepancies, working directly with clinical staff.
Ongoing
UAE · Freelance

Freelance — Tech, Data & Content

Early-stage UAE startups

  • Websites, data preprocessing, Power BI reporting, and brand content (Adobe Express, Canva) for early-stage clients.
Remote
Bangalore · Internship

Project Intern — Data Science & Visualization

Skill Vertex

  • Built interactive dashboards and visual reports in Power BI and Matplotlib, with a focus on data storytelling.
  • Recognised by the tech team for proactive communication.
Projects

Other things I've built

Separate from Kognition, built during certifications and on my own.

LLM · Retrieval

RAG Chatbot on Oracle Cloud (OCI)

A domain-specific chatbot built during the OCI Generative AI certification. A retrieval-augmented pipeline with document ingestion, vector embeddings, and similarity search, served by Python backend services over REST APIs.

OCI GenAIPythonEmbeddingsVector searchREST
Generative AI · Web

AI Image Generator

A self-directed web app that generates images from natural-language prompts through the OpenAI DALL·E 3 API, with API keys handled server-side.

PythonStreamlitDALL·E 3
BI · Visualization

Power BI Dashboards

Interactive dashboards and KPI reporting from internship, freelance, and course work, focused on turning raw operational data into something a non-analyst can act on.

Power BIDAXExcelMatplotlib
Agentic AI

NOLAN — Network Operations & Logic Analysis Node

My personal AI assistant: a voice-controlled system that runs entirely on my own machine. Built in Python's standard library with no frameworks or dependencies, including a custom PDF parser, spreadsheet reader, and a canvas force-directed knowledge graph over my own documents. It speaks both directions through ElevenLabs, reasons with Claude Opus 5, and reads live infrastructure state from Railway, GitHub and Twilio to answer questions about Kognition's pipeline in plain language.

Vanilla JavaScript + Canvas2DClaude Opus 5.5Py Standard LibraryRailway GraphQL API
Toolkit

Skills

Programming & Data Science

PythonSQLPandasNumPyStatisticsEDAData cleaningMachine learningVisualization

AI & Generative AI

LLMsRAGPrompt engineeringAgentic workflowsEmbeddingsVector searchClaude APIOpenAI API

BI & Analytics

Power BIAdvanced ExcelTableauKPI reportingForecastingAnomaly detectionAutomated reporting

Cloud & Engineering

OCIAWSGCPRailwaySupabaseREST APIsGit / GitHubCI/CDFlask

Web

HTML5CSS3JavaScriptReactStreamlitNetlify
Certifications

Certifications

  • Machine Learning SpecializationStanford · DeepLearning.AI
  • OCI Generative AI ProfessionalOracle
  • Trustworthy Generative AIVanderbilt · Coursera
  • Cloud Computing with AWSEngineer Core · IIIT Allahabad
  • Building ML Solutions with Azure DatabricksMicrosoft
  • Programming with Generative AI & CopilotIIT Guwahati
  • Claude 101 · AI Fluency & FrameworksAnthropic Academy
  • SQL DeveloperInfosys Springboard
  • Excel Dashboards & Data AnalysisUdemy
  • Certified Python ProgrammerIPCS Global
  • Microsoft Power BI CourseworkOffice Masters
  • PL-300 Power BI Data Analyst AssociateIn progress
  • Google Professional ML EngineerIn progress
Education

Education

Undergraduate

BS (Hons) Data Science & Artificial Intelligence Engineering

Indian Institute of Technology Guwahati

Junior College

Computer Science

St. Michael's Anglo Indian Higher Secondary, Kerala