Abu Dhabi · UAE

I architect AI systems
that ship.

Cut document research time from 2 hours to 10 seconds.

Built an AI chatbot now handling 18K+ queries a month.

Made 200+ non-technical staff fluent in SQL — without SQL.

Shipped a vision pipeline saving 2,000 staff hours every month.

Reduced GPT-4 API costs by 38% while improving response quality.

Ask my AI anything

0+
Years Shipping Software
0+
Production AI Systems
0K+
Documents Processed
0K+
Monthly AI Queries
0%
Research Time Saved
0K+ hrs
Staff Hours Saved / Mo

Shipped systems for · Trained by

SCAD

Statistics Centre Abu Dhabi

MoHRE

Ministry of Human Resources · UAE

Microsoft AI Developer Program

Trained · Aug–Sep 2025

15+ Years

Shipping production software

01About

Fifteen years of shipping software. The last three building AI.

I'm Mazhar — an AI Solutions Architect in Abu Dhabi. Before LLMs were cool I was shipping .NET systems for governments and ministries serving millions of users. When the world changed, I went deep on retrieval, prompt engineering, evals, and the unglamorous plumbing that makes AI work in production.

I'm a vibe-coding advocate — shipping fast, end-to-end, refusing to confuse a demo for a product. The systems I build run every day, in front of thousands of users, and the metrics are public.

Lesson · 01

A demo is a hypothesis. Production is the only evidence.

— Lesson learned

02Selected Work

Three systems. In production.

012025SCAD

Enterprise RAG Document Intelligence System

Challenge

100K+ government documents scattered across legacy systems with no unified search.

Approach

Architected end-to-end RAG pipeline using Azure OpenAI and Cognitive Search with hybrid retrieval and re-ranking.

  • Reduced information retrieval time from 2–3 hours to under 10 seconds
  • Achieved 92% accuracy on complex multi-document queries
  • Processing 5K+ queries monthly with 87% user satisfaction
  • Cut document research costs by 65% through automation
10s
time
92%
accuracy
-65%
cost
5K+/mo
queries
GPT-4Azure Cognitive SearchLangChainPineconeAzure Functions.NET CoreAngular

rag.pipeline

● live
CORPUSINDEXRETRIEVEGENERATE · CITE100K+ DOCSdoc_001.pdfdoc_002.pdfdoc_003.pdfdoc_004.pdfdoc_005.pdfdoc_006.pdfAR + ENpolicy · stats36 chunks liveChunksemantic · 300 tokEmbedgemini · 3072dPINECONEserverless · freeUSER QUERY"How did you cutcosts by 65%?"Hybrid RetrieverBM25 + vector · RRFRe-rankcohere · cross-encTOP-5 CHUNKS[1] case-rag-decisions 0.92[2] stack-cost 0.87[3] case-rag-timeline 0.81Groq · Llama 3.3 70Bstreaming · <800msANSWER + CITATIONSWe compressed context [1],filtered chunks via rerank[2], and used few-shot [3]to cut tokens 65%.↻ eval harness · 200-question gold set runs on every change92% ACCURACY · 65% COST CUT · 5K+ QUERIES/MO · <2s END-TO-END
022023 – PresentSCAD

Intelligent Conversational Analytics Platform

Challenge

Non-technical users needed SQL database access without coding knowledge.

Approach

Built natural language to SQL query system with conversational interface and automatic error correction.

  • Enabled 200+ non-technical staff to query databases using plain English
  • Handles 15K+ queries monthly across 8 different databases
  • Reduced analytics request backlog by 70%
  • 85% query accuracy with automatic error correction
200+
users
15K+/mo
queries
-70%
backlog
85%
accuracy
GPT-4Semantic KernelAzure OpenAI.NET Core Web APIAngularSQL Server

nl_to_sql.flow

● live
INPUTREASONEXECUTEVALIDATE · REPAIR · DELIVERUSER QUESTION"Top 5 sectorsby employmentin 2024?"EN · AR · multilingualIntent + Schemasemantic-kernelSQL SynthesisGPT-4 · 80 few-shots8 DATABASES · READ-ONLYcensus_2024labor_forcetrade_statsdemographicsGENERATED SQLSELECT sector, SUM(emp) AS totalFROM census_2024WHERE year = 2024GROUP BY sectorORDER BY total DESC LIMIT 5;Validate + Repairexecution-awareRESULT SET5 rows · 47 ms→ chart + explanation↻ +13% accuracy from repair loop200+ NON-TECH USERS · 18K+ QUERIES/MO · 85% ACCURACY · ROW-LEVEL SECURITY
032025SCAD

Document Processing & Vision AI Pipeline

Challenge

Manual processing of 1000+ daily documents (PDFs, scanned images, forms).

Approach

Built intelligent document processing pipeline using GPT-4 Vision and Azure Form Recognizer.

  • Automated 80% of document classification and data extraction tasks
  • Reduced processing time from 15 minutes to 30 seconds per document
  • 94% accuracy on structured form extraction
  • Saved 2000+ staff hours monthly
80%
automation
30s
time
94%
accuracy
2K+/mo hrs
saved
GPT-4 VisionAzure Form RecognizerAzure FunctionsBlob StorageCosmos DB

vision.pipeline

● live
INTAKEROUTEEXTRACT · MERGEVALIDATE · DELIVERINBOX · 1K/DAYPDF420JPG180PNG150FORM210SCAN40AR + EN contenthandwriting · tables14 document typesClassifierroute by typeForm Recognizerstructured formsGPT-4 Visionunstructured · ARTesseract · OCRlegacy scansMerge + Enrichper-field confidenceSTRUCTURED · COSMOS DB{ type, fields, conf } emp_id: "12345" (0.99) date: "2024-01" (0.94) notes: ar+en (0.71)↳ if conf < 0.85 → human review queue80% AUTOMATED · 94% EXTRACTION · 2K+ HRS SAVED/MO · 30s VS 15min · 14 DOC TYPES

Also shipped

2023

AI Conversational Chatbot (18K+ Monthly Queries)

SCAD — Statistics Centre Abu Dhabi

Deployed conversational AI chatbot with context-aware multi-turn dialogue and intent routing.

Azure Bot FrameworkGPT-4Azure OpenAILangChain.NET Core
2025

Arabic ↔ English Statistical Report Translator

SCAD — Statistics Centre Abu Dhabi

Built a GPT-4 translation copilot with a SCAD-specific glossary (UN SDG terms, demographic taxonomies) and a side-by-side reviewer UI.

GPT-4Azure OpenAICustom GlossaryAngular
2025

Smart Meeting Summariser & Action-Item Extractor

SCAD — Statistics Centre Abu Dhabi

Pipeline ingests Teams transcripts, summarises in Arabic + English, extracts owners and deadlines, and posts structured items to Planner.

WhisperGPT-4Microsoft GraphPower Automate
2024

Internal Policy & HR Q&A Bot

SCAD — Statistics Centre Abu Dhabi

Slim RAG pipeline over the policy library with citation-first answers and an escalation hand-off to a human HR contact.

Azure OpenAIAzure AI SearchLangChain.NET Core
2025

Inbox Triage & Auto-Reply Drafting Copilot

Internal use

Outlook add-in that classifies incoming mail by intent, drafts a tone-matched reply, and learns from accept/reject signals.

GPT-4Microsoft GraphOffice Add-in.NET
2025

AI Code-Review Assistant for .NET Repos

Internal engineering

GitHub Action that posts inline review comments — security checks, async/await pitfalls, EF Core anti-patterns, and a prompt-injection scanner for any AI-touching files.

GPT-4GitHub ActionsRoslyn AnalyzerOctokit
2024

Statistical Data-Quality Anomaly Detector

SCAD — Statistics Centre Abu Dhabi

Hybrid stats + LLM pipeline: classical outlier detection flags suspect rows, then GPT-4 reasons about whether they're real changes vs. likely data-entry mistakes.

PythonGPT-4scikit-learnAzure Functions
2024

Open-Ended Survey Response Coder

SCAD — Statistics Centre Abu Dhabi

Few-shot Arabic-first LLM coder with confidence thresholds; high-confidence answers auto-coded, low-confidence ones routed to a human-in-the-loop UI.

GPT-4Azure OpenAIArabic NLPAngular
2024

Smart Form Validator (Arabic Address & Name Parsing)

SCAD — Statistics Centre Abu Dhabi

Light GPT-3.5 normaliser + transliteration model that standardises names, splits address components, and validates them against the national address registry.

GPT-3.5Azure FunctionsArabic Transliteration
2024

Knowledge-Base Auto-Tagger & Linker

SCAD — Statistics Centre Abu Dhabi

Batch embedding + classification pipeline that assigns SDG topics, year, language, and confidentiality tier, plus suggests related documents.

text-embedding-3GPT-4Azure AI Search.NET
2025

Executive Dashboard Narrator

SCAD — Statistics Centre Abu Dhabi

Scheduled job reads dashboard datasets, runs significance tests, and writes a 4-paragraph bilingual executive brief grounded in the actual numbers.

Power BIGPT-4Azure FunctionsLogic Apps
2022

Regulatory & Policy Change Watcher

MoHRE — UAE Government

Daily crawler + LLM diff summariser that posts only material policy changes — with citations — to a Teams channel.

PythonGPT-4Microsoft TeamsAzure Logic Apps
2021

Citizen Service Request Auto-Router

MoHRE — UAE Government

Intent + topic classifier on top of the request text routes to one of 22 specialist teams, with a confidence-gated escalation path.

BERT (Arabic)ML.NETAzure MLWeb API
2017

Real-Time Social Sentiment Engine

TRG Tech

Streaming pipeline classifies sentiment + topic, detects spikes, and pushes alerts to subscriber dashboards.

Node.jsRabbitMQSentiment AnalysisAngular
2014

Lease Application Lead Scoring

NETSOL Technologies

Logistic-regression scoring model on application + bureau features, exposed as an API into the leasing workflow.

R.NETSQL ServerREST API
2025

Prompt & RAG Evaluation Harness

SCAD — Statistics Centre Abu Dhabi

Open-source-flavoured eval harness with ~400 graded queries, judge-LLM scoring, and a CI step that blocks regressions in retrieval recall and answer quality.

GPT-4 (judge)PythonGitHub ActionsDVC
2026

Portfolio Model Context Protocol Server

Personal · Open-source

stdio MCP server exposing typed tools (get_projects, get_case_study, search_writing) so any compatible client can ground answers in real portfolio data.

MCPTypeScriptNode stdioZod

Lesson · 02

The bottleneck is never the model — it's chunking, retrieval, and prompts.

— Lesson learned

03Stack

The toolkit behind the systems.

01

Generative AI & LLMs

Building with frontier models in production.

Green items link to proof — a project or article where I used them

02

RAG & Vector Search

Retrieval pipelines that scale to millions of docs.

Multi-stage RetrievalHybrid SearchRe-rankingPineconeFAISSChromaAzure Cognitive SearchWeaviateOpenAI EmbeddingsSentence Transformers

Green items link to proof — a project or article where I used them

03

AI Frameworks

Orchestrating agents, tools, and memory.

LangChainSemantic KernelLlamaIndexHaystackAzure Bot FrameworkMLflowAzure ML

Green items link to proof — a project or article where I used them

04

Cloud & Architecture

Secure, event-driven, cloud-native platforms.

Green items link to proof — a project or article where I used them

05

Full Stack

End-to-end systems from API to UI.

.NET Core 8ASP.NET Web APIAngular 17ReactTypeScriptNode.jsSQL ServerRedis CacheCI/CD Pipelines

Green items link to proof — a project or article where I used them

06

Conversational AI & NLP

Natural dialogue that resolves real problems.

Multi-turn DialogueIntent ClassificationEntity ExtractionSentiment AnalysisDocument UnderstandingSummarizationQ&A Systems

Green items link to proof — a project or article where I used them

04Experience

git log --author="mazhar"

Nov 2022 – Present

Abu Dhabi, UAE

Senior System Analyst / AI Platforms

@ Statistics Centre — Abu Dhabi (SCAD)

  • +Architected enterprise RAG system processing 100K+ documents using GPT-4 and Azure Cognitive Search, reducing research time by 95%
  • +Deployed conversational AI chatbot handling 18K+ monthly queries with 90% first-contact resolution, cutting support costs by 43%
  • +Built natural language SQL query interface enabling non-technical users to access 8 databases, processing 15K+ queries monthly
  • +Designed prompt engineering framework reducing GPT-4 API costs by 38% while improving response quality by 15%
  • +Led migration of 8 legacy monolithic applications to AI-enhanced microservices
GPT-4Azure OpenAILangChainSemantic KernelPinecone.NET Core 8Angular 17Kubernetes

July 2018 – Nov 2022

UAE

Senior Full Stack Engineer / Team Lead

@ Ministry of Human Resources & Emiratisation (MoHRE)

  • +Led development of Tasheel Systems — 50+ labor and HR services applications serving 2M+ users annually
  • +Modernized legacy codebase to microservices architecture, improving performance by 45%
  • +Designed RESTful APIs consumed by 30+ internal and external systems with OAuth 2.0
  • +Implemented Redis caching strategy reducing database load by 55%
  • +Led team of 6 developers using Agile/Scrum with 95%+ sprint completion rate
.NET Core 5/6Angular 12-14ReactDockerKubernetesAzure DevOpsRedisAWS

June 2015 – June 2018

Lahore, Pakistan

Senior Software Engineer / Technical Lead

@ TRG Tech

  • +Led cross-functional team of 8 developers (Full Stack, iOS, Android)
  • +Built real-time sentiment analysis engine processing 100K+ social media posts daily
  • +Developed social media monitoring platform integrating Twitter, Facebook, Instagram APIs
  • +Architected data pipelines processing 1M+ records daily for business intelligence
.NET Framework 4.6AngularNode.jsSQL ServerSocial Media APIsSentiment Analysis

Dec 2012 – June 2015

Lahore, Pakistan

Software Developer

@ NETSOL Technologies

  • +Maintained and enhanced large-scale financial leasing suite for international clients (FIAT, CNH Industrial)
  • +Delivered 20+ features for enterprise financial management system
  • +Reduced bug count by 35% through code refactoring and unit testing
.NET Framework 4.5ASP.NET MVCAngularJSSQL ServerCrystal Reports
05Training & Programs

Trained by Microsoft.

Microsoft Official Course completions and learning paths. Exams AI-102 / AZ-305 not yet attempted — the knowledge is applied daily in production at SCAD.

Training completed · exam not yet attempted

Trained

Azure AI Engineer track (AI-102)

Worked through the full AI-102 curriculum — Azure OpenAI, Cognitive Services, knowledge mining, conversational AI — applied directly in production at SCAD.

Microsoft Learn

Training completed · exam not yet attempted

Trained

Azure Solutions Architect track (AZ-305)

Self-paced study of architecture design patterns for Azure, identity, governance, data platform, and business-continuity design.

Microsoft Learn

07Live Demo

This site runs RAG on itself.

The demo below queries a Pinecone vector index of my CV content — in real time. Type any question and watch the retrieval pipeline run: embed → retrieve → rank → generate with citations.

01

Embed query

Your question is converted to a 3072-dimension vector using Gemini Embedding 001 (free).

02

Retrieve + Rerank

Top 10 candidates fetched from Pinecone, then Cohere rerank-v3.5 reorders them by relevance.

03

Cited generation

Groq Llama 3.3 70B synthesises a cited answer from the top 5 chunks — in ~1 second.

16 knowledge chunks · Pinecone serverless · Gemini embed (3072d) · Cohere rerank-v3.5 · Groq Llama 3.3 70B · All free tier.

Live RAG Demo

Pinecone · Gemini embed · Cohere rerank · Groq Llama 3.3

Live

New · Agent-native CV

I shipped my CV as an MCP server.

If you use Claude Desktop, Cursor, or any MCP-aware client, point it at this repo and your agent can interview my CV directly — six structured tools, two resources, zero hallucination.

Wire it up

Lesson · 03

Measure or it didn't happen. Latency, accuracy, cost — define them first.

— Lesson learned

06What colleagues say

Words from the people who shipped with me.

Mazhar architected our RAG document intelligence system from scratch — 100K+ government documents, retrievable in seconds. His ability to translate a vague business problem into a precise, production-ready AI architecture is rare. He delivered on time, measured everything, and the system has run without issues for over a year.

Senior Director

Digital Transformation · Statistics Centre Abu Dhabi (SCAD)

Direct stakeholder

He built the NL-to-SQL analytics platform that changed how 200+ of our non-technical staff work. What impressed me most wasn't the technology — it was his insistence on measuring accuracy before and after every change. He doesn't ship until the numbers say it's ready.

Head of Analytics

Data & Analytics · SCAD

Internal client

Mazhar led the backend architecture for Tasheel — one of the highest-traffic government service platforms in the UAE. He brought the kind of calm, systematic thinking that made a complex distributed system feel simple. His code reviews alone upskilled the entire team.

Engineering Manager

Platform Engineering · MoHRE UAE

Direct manager

* Quotes represent the substance of feedback received. Names withheld at request of colleagues; full references available on request.

08Writing

Patterns from production AI.

Nov 2025

12 min read

RAG from prototype to production: what nobody tells you

Chunking strategies, re-ranking, hybrid search, eval frameworks — the six decisions that separate a demo from a system that runs 24/7 in front of thousands of users.

RAGLangChainAzure OpenAIProduction

Sep 2025

9 min read

Prompt engineering patterns I actually use in production

Few-shot, chain-of-thought, function calling, and system-prompt hygiene — with real examples from the systems I've shipped and the cost/quality trade-offs of each.

GPT-4Prompt EngineeringAzure OpenAI

Jul 2025

10 min read

Getting NL-to-SQL to 85%+ accuracy without fine-tuning

How schema injection, intent classification, execution-aware repair loops, and a good evaluation harness got our conversational analytics platform to production-grade accuracy.

Semantic KernelSQLGPT-4NLP

May 2025

7 min read

Cutting GPT-4 API costs 38% without hurting quality

The prompt engineering framework we built at SCAD — caching, token budgeting, model routing, and eval-driven iteration — that saved tens of thousands annually.

Cost OptimisationGPT-4LLMOps

Articles publishing soon. Follow on LinkedIn to be notified.

Stay in the loop

New articles & shipped projects — straight to your inbox.

~1 email a month. Real lessons from production AI. No spam, no recycled LinkedIn content.

09Contact

Let's build something that ships.

Open to senior AI architecture roles, consulting engagements, and speaking opportunities — based in Abu Dhabi, working globally.

Start a conversationDownload CVDOCX