mountaindata.ai

Aman Gupta · Data engineer · BI · Consultant

Your data, gathered. Your questions, answered.

Pulled from the tools you already use into a database you own, cleaned, and put on a dashboard your team opens every morning.

Let's talkSee the four offers

Solutions

  1. 01 · Most teams start here

    Your data, somewhere you own

    From one tool you use, into a database you own, updated automatically.

    • From one system: your database, or an app your team already uses
    • Into a warehouse you own
    • Updated on a schedule, loading only what changed
    • Runs where you want it, and you keep the keys
    • You hear about it when it fails

    from $560 (2 days)

    2 to 5 days of work, over 1 to 2 weeks

  2. 02

    Ready for questions

    The same data, cleaned and organised so questions have answers.

    • Everything in 01, plus
    • Cleaned and combined into tables that match how your business works
    • Checked every update, so bad data is caught before it reaches a report
    • Named for your questions, not the software's

    from $1,400 (5 days)

    5 to 10 days of work, over 2 to 3 weeks

  3. 03

    On a dashboard

    All of the above, on a screen your team opens every morning.

    • Everything in 02, plus
    • One or more dashboards, built to your brief
    • A warning when the numbers are stale
    • A walkthrough for whoever will use it

    from $2,240 (8 days)

    8 to 15 days of work, over 3 to 5 weeks

  4. 04

    Something else

    For what does not fit the first three.

    • Migrations: moving off Fivetran, Stitch or Airbyte
    • Adding more systems to what you already have
    • Reverse ETL: sending data back out, to HubSpot, Sheets or your own app

    Let's discuss

Fixes for 30 days after handover are included. Support by the month after that, if you want it.

In production at Navit since 2023 · 145 merged pull requests to dlt · 32 posts for the dlt blog. See the work →

Sources
  • HubSpot
  • Salesforce
  • Stripe
  • Shopify
  • Zendesk
  • Google Analytics
  • Google Sheets
  • Notion
  • Postgres
  • MySQL
  • MongoDB
+19 more

Also any REST API, and files in S3 or Google Cloud Storage.

Destinations
  • Snowflake
  • BigQuery
  • Databricks
  • Redshift
  • ClickHouse
  • Postgres
  • DuckDB
  • MotherDuck
+20 more

Also MS SQL, Iceberg, Delta Lake and files in a bucket.

Built with

dlt for loading, incremental or full refresh; dbt or dlt for the models; scheduled on GitHub Actions, Airflow or dltHub; Metabase or Looker Studio for dashboards.

How it works

  1. 01

    A call

    Half an hour on what you have and what you want to see.

  2. 02

    A written offer

    What is in, what is out, the range and the calendar.

  3. 03

    The build

    A short check-in each week. Nothing switches off until the new thing is proven.

  4. 04

    Handover

    Documentation, access and a walkthrough. You own all of it.

Not sure which one? Start with the call.

Half an hour, no preparation needed. You leave knowing which of the four fits, and the written offer follows.

Let's talkOr email

Experience

Data Engineersince2023

dltHub

145 merged pull requests to the open-source dlt framework: documentation across dozens of sources and destinations, benchmark studies, and Fivetran/Stitch-to-dlt migration engineering.

Data work at dltHub: documentation, migrations and benchmarks

Data Engineeringsince2023

Navit

Pipelines from Postgres, HubSpot and Freshdesk into BigQuery through bronze, silver and gold dlt layers, orchestrated on dltHub, served in Metabase dashboards.

Navit's data stack: three years in production, then its move to dltHub

Data Architect & BI2023 – 2025

RentLondonFlat.com

Schema design for the in-house app, data quality checks on its database, incremental pipelines into BigQuery, and property and customer scores, shown in Looker.

Data work at RentLondonFlat.com: from app schema to dashboards

Scientific Officer2014 – 2023

Bhabha Atomic Research Centre

Nine years in capital projects: cost models and rate analysis, tender evaluation, and billing reconciliation across budgets of $50M+.

For nine years, my data had to be good enough for a judge