AI & Data Product Leader · Builder

I build AI products on trusted data foundations.

I turn complex data, AI, and regulatory challenges into products people can understand, trust, and use.

Portrait of Ankit Aggarwal
15+ years
making complex data useful.
15+years in data
and product
50%fewer audit
findings
21d → 6htime to value
for new products
250K+users reached
through GenAI
About me

Builder at heart.

I don’t wait for a perfect start. I just get started.

Right after school, I took my first job at McDonald’s for ₹2,000 a month, about $22. I studied computer science at Thiruvalluvar University while working, interned at HCL Tech, and broke into core tech at iQor in 2009. Since then, I’ve spent 15+ years building with data, and today I lead data and AI products at JPMorganChase.

What I bring to data

Most of my career has been in regulated financial data, where accuracy is non-negotiable and every number needs a traceable origin. That work has given me depth across the full data lifecycle:

  • Data modeling and integration. Schema and ETL specs, transformation logic, and SQL, from source systems to the products that depend on them.
  • Lineage, quality, and governance. I designed the firmwide lineage platform behind BCBS 239 compliance, which cut audit findings by 50%.
  • Semantics and knowledge graphs. Ontologies and graph databases (Neo4j) that give data meaning, so people and AI systems can interpret it correctly.
  • Data access and entitlements. API specifications and permission models, including attribute-based access control, so the right people see the right data.
  • Modern data platforms. Python, SQL, Databricks, and Snowflake, applied to enterprise products used by internal and external clients.
  • AI on top of data. RAG and vector search, agentic patterns (tool use, MCP, multi-agent orchestration), prompt engineering, and LLM evaluation.

I’ve also learned that data work succeeds or fails on people. I’ve run client workshops, given talks and panels, and taken platforms through UAT, phased rollouts, and production support.

How I work

  1. Understandable before clever. If people can’t explain a data model, they won’t trust it.
  2. Trust is the product. In regulated data, every number needs a traceable origin. In AI, every answer needs grounding.
  3. Build to learn. I prototype myself, because the hardest data problems only show up once you build.
  4. Show the work. That’s why this site exists.
02 / Core Capabilities
01

Data Product Management

Self-serve platforms and reusable patterns that move data from strategy to adoption.

02

Build & Ship Large Enterprise Softwares

Building scalable enterprise software from product strategy through launch and adoption.

03

Forward Deployed Product Manager

Partnering with customers and engineering teams to turn complex needs into shipped product capabilities.

03 / Selected work

Ideas, shipped.

Some of the platforms and systems I've helped bring into the world.

01Startup

Graph Context Layer

Graph-powered context, ontologies, and semantic models that make data intelligible.

Knowledge graphsOntologiesContext
02JPMorganChase

Data Product Builder

Owned the vision and roadmap for a platform launched from zero to external clients, now growing monthly active users by roughly 100% year over year.

Product strategyData platformsAdoption
03JPMorganChase

Fusion By J.P.Morgan

Built world class data management enterprise platform for external and internal clients by leveraging graph database.

Data managementGovernanceEnterprise
04JPMorganChase

Data Lineage for BCBS239

Designed and built firmwide platform that harness data lineage, quality monitoring, remediation, semantic context, cutting audit findings by 50% using.

LineageData qualityRisk
05Startup

No code SaaS Enterprise Platform

Built Mezocliq's low-code data platform and its metadata, profiling, quality, and lineage layer so citizen developers could move from ingest to insight.

0 to 1No-codeData products
06Startup

Attribute Based Access Control

An authorization model that grants or restricts access by evaluating real-time attributes of the user, resource, operation, and environment instead of using static roles

LineageData QualityData GovernanceData Management
04 / Experience

A career in useful complexity.

View Details
2021 - now

JPMorganChase

Executive Director, Technical Product Management

Enterprise data platforms, governance, and data quality.
2014 - 2021

meZocliq

Director, Technical Product Management

Self-serve and no-code data management platform.
2006 - 2013

iQor

Product Manager, Data & Reporting Products

Data integrations and internal reporting platform.
2005 - 2006

HCL Tech

Intern - Software Development

Wrote my first line of code for US mortgage firms.
2005

mcDonalds

Crew Member

Made all kinds of burgers.
05 / Writing

Thinking in public
about data.

Notes on data products, context, governance, and what it takes to make complex systems useful.

Read on Substack
06 / Education

Education

MIT Professional Education

Product

Institute of Management Technology

MBA

Thiruvalluvar University

Bachelors in Computer Science

07 / Start a conversation

Have a data problem
worth untangling?

Whether you're building a platform, navigating regulation, or looking for a clearer way through complexity, I'd like to hear from you.