Background

AI product manager with four years of enterprise data delivery behind it

I started in enterprise data execution: workflows, requirements, releases, QA collaboration, modernization, and stakeholder alignment. That work showed me where products break down and where a decision made earlier would have changed the outcome. Today I own product decisions on AI and data products, write the documents teams build from, and prototype with code when a diagram will not answer the question.

View Experience→Read the case studies→

Name: Deepankar Sharma
Title: AI Product Manager · Data and AI Products
Location: Kolkata, India · Remote, US and EU overlap
Experience: Around 4 years in enterprise data
Focus: AI products, data products, execution-ready documentation

02 · Origin Story

How Technical Execution Shaped My Product Thinking

I spent the first part of my career close to data workflows, requirements, defects, releases, and stakeholder needs. The view from there is uncomfortable but useful: you see exactly where unclear problems turn into rework, where weak framing leads to scope drift, and where missing documentation forces teams to re-derive context every sprint.

That experience is what eventually became my Problem to Product Workflow. The 10 steps did not start as a framework. They started as the questions I kept wishing someone had answered before the work reached engineering.

Insight 01

Unclear problems create execution waste

Most engineering rework I saw was downstream of a problem that was never properly defined. Step 1 of the workflow exists because of this.

Insight 02

Strong documentation improves alignment

PRDs, acceptance criteria, and Definition of Done are not paperwork. They are the cheapest way to remove ambiguity before sprint commitment.

Insight 03

Product decisions get stronger when grounded in technical reality

Feasibility, data readiness, and architectural runway should sit inside the product conversation, not after it. Step 5 was built for this.

03 · Experience

Experience Across Data Products, Delivery, and Client Work

Nov 2023 to Present· Kolkata, India

Technical Product Owner

Cognizant

Enterprise Data Products · Salesforce to Snowflake

Owned end-to-end delivery of an enterprise Salesforce-to-Snowflake data product. Translated business intent into a prioritized backlog, aligned sales, analytics, and architecture stakeholders, and used GenAI-assisted rapid prototyping to validate Snowflake architectures and CTE-based transformation logic before sprint commitment. Also led discovery and feasibility framing for Informatica-to-Snowflake modernization.

Aligned sales, analytics, and architecture stakeholders on requirements and acceptance criteria
Translated business intent into a prioritized backlog of 80 plus user stories across two squads of 6 to 10 members
Reduced manual operations by around 40 percent and increased analyst trust in downstream reporting
Improved sprint predictability by around 25 percent and reduced defect leakage by around 20 percent
Partnered with architecture and QA to convert legacy mapping logic into a supportable target state
Improved pipeline performance by around 30 percent and reliability by around 35 percent
Drove alignment through sprint ceremonies, demos, and release readiness reviews

Backlog OwnershipSalesforce-to-SnowflakeInformatica ModernizationAcceptance CriteriaGenAI PrototypingAI-Led DevRelease Readiness

Aug 2022 to Oct 2023· Kolkata, India

Developer

Cognizant

Analytics-Facing Data Workflows

Supported product delivery for analytics-facing data workflows. Partnered with senior owners on requirement clarification, defect triage, and stakeholder updates. Designed reusable validation routines as a productized data-quality capability. Built early fluency in stakeholder communication, sprint coordination, and lifecycle documentation.

Improved end-to-end workflow efficiency by around 20 percent
Designed reusable validation routines as a productized data-quality capability
Reduced manual verification effort by around 25 percent
Delivered 10 plus release-ready features across Agile sprints
Built early fluency in stakeholder communication, sprint coordination, and lifecycle documentation

Requirement ClarificationDefect TriageValidation RoutinesSprint CoordinationStakeholder Updates

Live client work· Independent

Independent Product Consultant

SimpliLEAD

Company Site and Product Hub for a HI-EI-AI Research Company

Worked with SimpliLEAD, a UAE-based HI-EI-AI ontology, research, and product development company, to build the site that anchors its ten consulting portfolios and its flagship HiMirrorly and HiCoachly product line. Shaped positioning, information architecture, and page scope, and used AI-assisted tools to draft copy and scaffold pages.

Positioned the HI-EI-AI ontology as the spine connecting ten consulting portfolios and the product line
Designed the information architecture for a company site serving enterprise buyers across the UAE, Saudi Arabia, Qatar, Oman, Bahrain, and Kuwait
Gave the flagship HiMirrorly and HiCoachly products a dedicated slot above the consulting portfolios
Tagged inbound inquiries by portfolio through a structured contact form
Applied product thinking outside enterprise constraints, in a small-team, independent setting

PositioningInformation ArchitectureAI-Assisted BuildPortfolio StructuringIndependent Client Work

04 · Selected Impact

Impact Across Product, Data, and Execution

A short list of measurable outcomes I helped deliver. All figures are honest, approximate, and tied to the role they happened in.

~40%

Reduction in manual operations on the enterprise Salesforce-to-Snowflake data product.

Backlog ownership · Acceptance criteria

80+

User stories translated from business intent across two squads of 6 to 10 members.

Requirement analysis · Story writing

~25%

Improvement in sprint predictability through structured refinement and acceptance reviews.

Agile execution · Definition of Ready

~20%

Reduction in defect leakage after introducing structured refinement and review.

QA alignment · Acceptance criteria

~30%

Pipeline performance improvement during Informatica-to-Snowflake modernization.

Feasibility framing · Data architecture

~35%

Reliability improvement during the same modernization effort.

NFR framing · Architecture alignment

10+

Release-ready features delivered as Developer across Agile sprints.

Sprint coordination · Lifecycle documentation

~25%

Reduction in manual verification effort through reusable validation routines.

Productized data quality · Reuse

05 · Skill System

What My Background Helps Me Do

Product Ownership

Discovery supportRequirement analysisBacklog ownershipUser storiesAcceptance criteriaStakeholder alignmentRelease readinessRoadmapping

Agile Execution

Sprint planningBacklog groomingSprint demosQA alignmentDefect triageRelease coordinationCross-functional delivery

Data Product Fluency

SnowflakeSalesforce-to-Snowflake workflowsInformatica modernizationSQLPythonMatillion ETLAirflowAWS S3 and Glue

AI-Assisted Product Execution

AI-Led DevelopmentGenAI-assisted prototypingLLM-powered discoveryArchitecture pressure-testingTechnical feasibility validationPrototype scaffoldingAI-assisted documentation drafts

06 · Operating System

The Operating System That Came From This Experience

The experience above is where my Problem to Product Workflow comes from. It is the same decision and evidence system behind every live product I work on, and it is the answer to the question I kept asking on delivery teams: who decided this, with what evidence, and when do we review whether it worked?

Step 01

Problem Discovery

Step 02

User and Segment Understanding

Step 03

Context and Market Research

Step 04

Product Strategy and Opportunity Hypothesis

Step 05

Feasibility, Data, and Risk Check

Step 06

MVP Scoping and Prioritization

Step 07

Experience and Solution Design

Step 08

Product Documentation

Step 09

Execution, Launch, and Monitoring

Step 10

Learning Review and Next Decision

Explore My Workflow→

07 · Client Work

Applying Product Thinking Beyond Enterprise Work

Client Work· Independent engagement

SimpliLEAD

I worked with SimpliLEAD, a UAE-based HI-EI-AI ontology, research, and product development company, on the site that anchors its ten consulting portfolios and its flagship HiMirrorly and HiCoachly product line. The engagement covered positioning, information architecture, and page scope across a broad consulting and product ecosystem, and applied the same discovery and documentation habits from enterprise work in a small-team, independent setting.

Read the case study→

What I contributed

Positioning and information architecture for ten consulting portfolios plus a flagship product line
Gave HiMirrorly and HiCoachly a dedicated flagship slot above the portfolio grid
Structured the contact form to tag inbound inquiries by portfolio
Applied product thinking outside enterprise constraints, in a small-team, independent setting

08 · Certifications

Certifications That Support My Product Direction

Microsoft AI Product Manager

2026

AI Product Management

Reinforces my direction toward AI product ownership: framing AI use cases, evaluation, trust, and lifecycle.

Certified Scrum Master, CSM

2025

Agile Delivery

Anchors the sprint, backlog, and ceremony discipline that runs through every role I have held.

Cisco CyberOps Associate

2024

Security and Operations

Supports the security and operational guardrails inside Step 5 feasibility and risk thinking.

Google Project Management Certificate

2023

Delivery Management

Backs the planning, coordination, and stakeholder discipline behind my delivery work.

Google Cloud Professional Data Engineer

2023

Cloud Data Architecture

Strengthens the data feasibility, pipeline reliability, and architecture conversations I run during discovery.

B.Sc. Physics Hons

Digboi College, 2020

Foundations

Built the analytical and first-principles habit that shows up in problem framing and evidence labelling.

09 · Execution Skill

AI-Led Development as a Product Execution Skill

I use AI-led development to move from product thinking to working prototypes faster. The goal is not just to generate code, but to validate feasibility, explore product flows, pressure-test technical decisions, and make product ideas concrete before committing full engineering effort. AI-led development supports product judgment. It does not replace it.

Feasibility

Feasibility Testing

Use rapid scaffolds to test whether an idea is buildable and operable before scoping a sprint.

Prototype

Prototype Scaffolding

Stand up flows, schemas, and UI shells fast enough to react to before they harden into stories.

Architecture

Architecture Pressure-Testing

Probe data models, transformations, and integration paths with real code, not just diagrams.

Translation

Documentation to Build Translation

Turn PRDs and acceptance criteria into concrete artifacts that surface the questions engineering will actually ask.

Learning

Faster Product Learning

Shorten the loop from idea to "is this real" so product learning compounds inside the same sprint.

10 · Closing

Why This Background Matters

My strength is the combination: technical fluency, data product execution, Agile delivery, stakeholder communication, PM documentation, AI-assisted prototyping, and product system thinking. Each one alone is common. The mix is what lets me turn an unclear problem into a backlog a team can actually ship, then back into a case study a recruiter or collaborator can actually evaluate.

If that combination fits what you are building or hiring for, the rest of this site is the proof of work.

See How This Experience Turns Into Product Work

Read the case studies, or email me about AI PM roles. I reply within one working day.

Explore My Workflow→Read the case studies→Email me→