dataonawsGet in touch
Offerings

Data & AI systems
in production.

Governed data platforms, agentic AI, shipped to production. Ten years hands-on, including several at AWS.

The approach

A few days on site. A written diagnosis.

1
Days 1-3

On-site discovery & first-impressions memo

Embedded with your teams: stack, code, data. Within 24h, the first-impressions memo — what we saw, what is solid, what needs attention, what we would do. In writing.

2
Week 1

Architecture proposal

Built on patterns proven in production. Costed, argued, trade-offs made explicit.

3
Then

Build, by the same team

The people who did the discovery do the build. Short fixed-scope engagements, seniors only, skill transfer included.

Offerings

Three families, one shared entry point.

Entry point

Discovery & first-impressions memo — a few days on site, a written diagnosis within 24h. Full data stack audit as an option.

Data platforms
  • Build & optimize your data platform on AWS state of the art, costs under control
  • Industrial-grade data pipelines use cases in production, not POCs
  • Migration & modernization to a modern data stack
  • Platform selection & architecture AWS native, Databricks, Snowflake
AI & agentic
  • End-to-end agentic AI use cases from business need to production, customer-facing included, with the ops to match
  • Scaling up agentic platform / agent factory, governance included, build vs buy
  • AI-assisted development Claude Code, Kiro, Codex on AWS, Amazon Quick, Cowork + governance framework
  • Talk-to-your-data governed chat over your lake / warehouse
Targeted expertise
  • On-demand AWS data expertise Redshift, Glue, EMR, Athena…
  • Data & AI-augmented contact centers Amazon Connect
  • Fractional data & AI leadership scoping, trade-offs, hiring
  • Training AWS cloud, data on AWS, AI on AWS, AI developer lifecycle
Reference patterns

Architectures proven before, adapted to your context.

Pattern 01

Governed chat-with-your-data

Query the warehouse in natural language without breaking access control: semantic layer, fine-grained permissions, reliability evals.

Bedrock · semantic layer · Redshift / Snowflake / Databricks
Pattern 02

Governed coding assistant for your teams

Claude Code, Cortex, Kiro, we guide you to the one that fits your needs and ship a deployment kit that keeps costs, fallbacks and data sovereignty under control.

Claude Code · Cortex · Kiro
Pattern 03

Modern data platform on AWS

The foundation that makes everything else possible: open lakehouse, modeling, quality, catalog, sized for a small team.

S3 / Iceberg · Glue · dbt · Lake Formation
Pattern 04

Agent-accelerated migration

Legacy (Oracle, SQL Server, Talend, DataStage…) to the modern data stack, with schemas, code and tests converted by agents and validated by humans.

conversion agents · dbt · automated testing

Each pattern is a documented starting point. The work is adapting it to your systems, constraints and teams.

Getting started

Your first deliverable lands
within the week.

Week 0

A 30-minute call to scope the discovery.

Days 1-3

On-site discovery, inside your teams, on your real stack.

24h later

The first-impressions memo, in writing. You decide what comes next with full information.

Book a discovery call →