DataBadger.aiDataBadger.ai
  • About
  • Work
  • Team
  • Contact
ENFR
Start a conversation
  • AboutWho we are, and how we build01
  • WorkEvery engagement, in detail02
  • TeamThe people on your project03
  • ContactStart with a problem, not a brief04
ENEnglishFRFrançais
Start a conversation
All projects

Blockchain gaming

A data and AI platform for a live game economy

Ingestion, a layered warehouse and an ML workspace over a live game economy — with a conversational layer that queries Snowflake directly through MCP.

Client

A blockchain game studio

0

SQL required to ask a question

A data and AI platform for a live game economy — Blockchain gaming

The challenge

What was in the way

A live game economy generates player, transaction and on-chain data continuously, and almost none of it is useful in the shape it arrives.

The studio needed those raw streams turned into something their team could ask questions of — without hiring a data team to do it.

The solution

What we built

The full analytics estate: ingestion pipelines on dlt, a layered dbt model over Snowflake, and a machine-learning workspace on the modelled data.

Over the warehouse we put a conversational layer that queries Snowflake directly through the Model Context Protocol, built on a multi-provider LLM factory so the studio was never locked to one vendor.

Because an analytics answer that is confidently wrong is worse than no answer, a supervisor pass verifies each result before it reaches the person who asked. The assistant meets the studio where they already work — a browser workspace for exploration, and a Slack bot for the questions that come up mid-conversation.

Key technologies

Warehouse
Snowflake with layered dbt models
Ingestion
dlt pipelines — player, transaction, on-chain
AI
MCP server over Snowflake; OpenAI, Anthropic, Bedrock, Gemini
Interface
Streamlit workspace, Slack bot, key-scoped sessions

Impact & results

01

Analytics in plain language

Game and economy data queryable by people who do not write SQL.

02

A modelled warehouse

Replacing ad-hoc extracts as the basis for every analysis.

03

No LLM vendor lock-in

Switching provider is a configuration change, not a rebuild.

04

Answers checked before they land

A verification pass sits between the model and the person asking.

Next project

An AI analyst for venture capital

Start here

Start with a problem, not a brief.

Tell us what is slow, what is manual, what is stalled, or what nobody understands any more. We will tell you honestly whether it is worth building.

Start a conversationinfo@databadger.ai
DataBadger.ai

Software, data and AI — studied, designed, built, deployed and maintained. End to end, by the people who architected it.

the technology department for companies that need one and do not want to build one.

www.databadger.ai — footer
$badger contact
emailinfo@databadger.ai
sitewww.databadger.ai
reply< 1 business day
$badger status
✓accepting new partners
✓est. 2022 · 100% client retention
✓12 engagements · 6 / 6 layers in-house
$badger sitemap
01aboutwho we are, and how we build
02workevery engagement, in detail
03teamthe people on your project
04contactstart with a problem, not a brief
$badger start --project
Start a conversation
$

© 2026 DataBadger.ai

Every partnership since 2022, still running.