Ayush Verma
Data & AI LeadAsk my AI twin
Ayush VermaAyush VermaData & AI Lead Dubai, UAE

Enterprise data governance, delivered

Data & AI Lead. Ten years in data and AI, the last eight in the UAE, mostly government and large commercial clients.

I currently lead the data governance programme for the Department of Government Enablement in Abu Dhabi, with a team of eight, taken over from Deloitte mid-flight. I still write the code as well as running the plan.

Ask my AI twinSee the career journey
AED 1.2BCampaign revenue · Noon
85%+Forecast accuracy · ADREC
5,800Data quality rules
40.5→100%Regulator score

01

Ten years, one question

What does someone actually have to decide, and can they trust the data enough to decide it? The first half of my career was the first question. The second half is the second.

2016–2023 · Decisions

Growth, campaigns and attribution at Noon, then a whole BI stack built alone at Right Farm. Analytics where the number on the line was revenue.

2023–2025 · Change

The data workstream through Silal’s acquisition of Safco: entity consolidation, master data alignment, reporting harmonised across two companies. What breaks when a business changes shape.

2025–now · Trust

Enterprise MDM and PMO governance at Contango, then a government data estate made defensible: catalogue, quality, lineage, ownership, and a regulator’s score taken from 40.5% to full conformance.

Most people who run governance have never carried a revenue number. Most growth analysts have never sat in front of a regulator. Doing both is the whole point.

Career Highlights

From analyst foundations to a government data estate

The same discipline scales from a single growth question to a governed estate feeding production AI.

$20M

Daily GMV of the Noon marketplace whose analyst team I led. Opportunity loss cut by 40% across the major campaigns.

10x

Revenue growth in four months at Right Farm. Inventory planning accuracy hit 90% and wastage fell four points.

150 bps

Gain in daily unit conversion at Namshi. From strategy reworked with the functional heads.

8 domains

Enterprise MDM on Profisee across 5 source systems. 30+ stewards onboarded and match accuracy up by about 20%.

15+

AI use cases defined for Q Mobility with 25+ stakeholders. Targeting a 15 to 20% efficiency gain.

3,646

Broken rule-to-column bindings repaired in a single day. By one Python sweep on the Purview API.

02

Capabilities

Six groups, listed exactly as they are worked. No aspirational entries.

01

Leadership

  • Programme and team leadership
  • PMO governance
  • Executive and C-suite reporting
  • Vendor transition and recovery
  • Risk and dependency management
02

Governance

  • Microsoft Purview (Data Map, Unified Catalog)
  • Profisee
  • Pimcore
  • Data quality
  • Critical data elements
  • Lineage
  • Stewardship models
  • Target operating model design
03

Regulatory

  • UAE PDPL
  • ADDA
  • NDMO
  • Classification and policy design
04

Platforms

  • Azure Databricks (Unity Catalog, Lakehouse)
  • Microsoft Fabric
  • Azure Synapse
  • SQL Server
  • ServiceNow
  • Google BigQuery
  • Power BI
  • Looker Studio
05

AI

  • Agentic frameworks
  • MCP
  • Copilot Studio
  • LLM profiling and anomaly detection
  • Multi-agent RAG
  • Computer vision
06

Engineering

  • Python
  • SQL (Databricks, Synapse, SQL Server, BigQuery)
  • REST API automation
  • R
  • Advanced Excel
  • Azure DevOps
  • Jira

03

Career journey

Ten years on one axis, most recent first.

Select a span to read the mandate behind it

Proportional timeline of 7 spans, from 2016 to the present. Bar length is time in role.

May 2026 – Present. Senior Manager, Data and AI, Delivery.

May 2026 – Present

Senior Manager, Data & AI (Delivery)

Publicis Sapient

Abu Dhabi

  • Lead a team of eight on the Department of Government Enablement's data governance programme. 844 datasets across 10 government verticals.
  • Took the programme over from Deloitte mid-flight; audited what had been built and offboarded the outgoing team's catalogue access.
  • Own the Purview build end to end: domains, data products, glossary, critical data elements, lineage and data quality. Sources are Azure Databricks, SQL Server, ServiceNow and Power BI.
  • Delivered 5,800 data quality rules. Took the regulator's score from 40.5% at first review to full conformance on several verticals.
  • Wrote Python automation on the Purview and Databricks APIs. One sweep repaired 3,646 broken rule-to-column bindings in a day.

Tap a span to read the mandate behind it

Proportional timeline of 7 spans, from 2016 to the present. Bar length is time in role.

May 2026 – Present

Senior Manager, Data & AI (Delivery)

Publicis Sapient

Abu Dhabi

  • Lead a team of eight on the Department of Government Enablement's data governance programme. 844 datasets across 10 government verticals.
  • Took the programme over from Deloitte mid-flight; audited what had been built and offboarded the outgoing team's catalogue access.
  • Own the Purview build end to end: domains, data products, glossary, critical data elements, lineage and data quality. Sources are Azure Databricks, SQL Server, ServiceNow and Power BI.
  • Delivered 5,800 data quality rules. Took the regulator's score from 40.5% at first review to full conformance on several verticals.
  • Wrote Python automation on the Purview and Databricks APIs. One sweep repaired 3,646 broken rule-to-column bindings in a day.

04

Ayush · AI twin
Ready

Ask my AI twin

Ask about the programmes, the numbers and the builds. It answers from my actual engagement history and shipped work — and it says so when it does not know.

Answers are generated from my profile context, so treat them as a summary rather than a record.

Transcript
AI twin replied:

Ask about the programmes, the numbers and the builds. I answer from my actual engagement history and shipped work, and I say so when I do not know.

05

Applied AI builds

Independent work, all self-built. Two are running in production.

In production

Verity EHS

A multi-agent RAG platform for chemical-sector environment, health and safety (EHS) compliance.

Multi-agent RAGCompliance
Read the case study
ayush-verma.com/ai-projects/verity-ehs
Verity EHS home screen: a regulation search bar with shortcuts to documents, permits and FAQ.
In production

LexDraft

A conversational AI agent that drafts legal documents.

Conversational agentDrafting
Read the case study
ayush-verma.com/ai-projects/prelegal
LexDraft landing screen: chat-first legal drafting with six document templates.

Sports analytics

A computer vision platform for football player performance.

YOLORTMPose-xByteTrack

In development

sports analytics — pose and track output
Sports analytics output: a tracked player with a pose skeleton overlaid on match footage, annotated with a coaching cue.

06

Credentials

Education, certifications, training and the rest of the record.

Education

IIT Roorkee — B.Tech, Engineering CGPA 8.2 / 10

Certifications

PMI Project Management Professional (PMP) Profisee End-to-End Fundamentals (LP100), 2025 Microsoft Purview: Cataloging & Governance, 2026

Training

IAPP AI Governance Professional (AIGP), BoK v2.1 Microsoft Copilot Studio: AI Agents, 2026 AI Coder: Agentic Engineering, 2026 Google Prompt Design in Vertex AI, 2025 Columbia Business School — Digital Strategies for Business Transformation, 2024

Also

English and Hindi UAE Golden Visa UAE driving licence National Mathematics Olympiad winner Grail Research Master Solver