DataAIChronicles

Consulting

Focused, senior help on the decisions that shape a data platform for years. I take on a small number of engagements alongside my work at GitLab, so each one gets my full attention.

Start a conversation

Where I can help

Data platform architecture review

For teams about to commit to a warehouse, lakehouse or catalog decision, or living with one that is not working.

What you get

  • A written review of your current platform against how it is actually used
  • Decision records for the two or three choices that matter most
  • A phased migration path your team can execute without me
Ask about this

AI cost and governance on Snowflake

For organisations adopting Cortex or other in-warehouse AI and wanting spend and access under control before the bill arrives.

What you get

  • Credit monitoring with thresholds and breach procedures
  • Role and OAuth design for AI workloads and MCP access
  • A cost model per team and use case your finance partner can read
Ask about this

Real-time CDC into Iceberg and Snowflake

For teams moving from batch extracts to change data capture and discovering how much more it changes than latency.

What you get

  • Source, transport and table design for CDC at your scale
  • Schema evolution, deletes and backfill strategy
  • Operational runbooks: replication slots, lag, recovery
Ask about this

Agents for data engineering

For platform teams letting coding agents touch pipelines and wanting deterministic guardrails around probabilistic tools.

What you get

  • Quality gates that apply the same rules to every agent and every human
  • A harness pattern for agents operating on your data stack
  • Evaluation and rollout plan for agent-authored changes
Ask about this

How an engagement runs

  1. A first conversation

    Thirty minutes on what you are building, what it runs on and what is getting in the way. No preparation needed.

  2. A written scope

    One page: the question we will answer, what you will receive and when. You decide whether to go ahead.

  3. The work

    Reviews, design sessions and written decisions, done alongside your team rather than handed over the wall.

  4. Handover

    Everything is written down so your team can carry it forward without me.

Track record

Outcomes from platforms I have designed and run at GitLab.

100%
of database changes captured by Project Siphon, closing a 30–40% gap
$100K+
a month in potential AI overruns prevented by Cortex governance
62%
pipeline runtime reduction, 1,456 compute hours saved a year
100 TB+
migrated from on-premise Kognitio to Snowflake on AWS