Shreyansh Dubey Data & Analytics Engineering
  • Data & Analytics Engineering Consultant
  • Microsoft Fabric · Power BI · SQL & Data Engineering

Shreyansh
Dubey

I take a business problem — not a ticket — and build the data platform, governed BI and engineering behind it.

  • 5.5+ years of client delivery
  • 17 engineers led across two concurrent engagements
  • ~200 enterprise users on a governed Fabric platform
  • US · UK client delivery, working remotely

Selected clients

    The practice

    Most analytics work fails at the translation, not the tooling.

    Five and a half years of client delivery across banking, wealth and asset management, travel, healthcare and manufacturing — usually as the person standing between the business question and the system that had to answer it. The engagements below start with what was wrong, not with what was built.

    1. 01

      Understand the problem first

      Requirements gathering with finance, operations and product leadership — then KPI definitions everyone has actually agreed to, before anything is modelled.

    2. 02

      Design the system, not the report

      Target-state architecture, lakehouse and warehouse design, and governed semantic models that hold their definitions so the estate stays trustworthy as it grows.

    3. 03

      Build what people decide with

      Executive and operational dashboards, funnel, retention and churn measurement, and financial analytics — built around the decision rather than around the shape of the data.

    4. 04

      Automate what shouldn't be manual

      Manual consolidation, paper workflows and third-party dependencies rebuilt as systems that run themselves on tooling the organisation already owns.

    More about how I work

    Selected work

    Four problems, and what changed

    All eight exhibits

    Consulting

    Where I'm brought in

    Start a conversation

    Working together

    Four ways to engage

    Start a conversation

    Rates depend on scope, duration and region, and are agreed in conversation rather than published here.

    Tell me what's not working.

    A short discovery conversation is usually enough to tell whether I'm the right person for the problem — and if I'm not, to point you at what is.