LinkedIn Ads ➡️ Snowflake, with 3-legged OAuth in the middle
LinkedIn's Marketing API needs a browser to authorize. Scheduled pipelines don't have browsers. I store the refresh token in Snowflake.
Latest
I migrated a Fivetran pipeline to dltHub for $0.65
One hour, five manual inputs, and $0.65 of tokens... enough to replace a daily Fivetran sync with a tested dltHub job
I needed 2,110 municipal contacts. So I built a crawler that refuses to guess.
A public contact list sounds easy... until 2,110 websites, four languages, changing staff, and source evidence get involved. 😅
Backing up Microsoft Planner with dlt and Azure Blob
Microsoft Planner has no native backup export, so I built a declarative dlt job that writes complete Graph snapshots to Azure Blob. 😎
My GTM container had 282 undocumented objects
I gave an agent a dependency graph to make it write notes on my GTM containers
Your coding agent keeps a diary
Every opencode session leaves a trail in a local SQLite file. dlt turns it into analytics you actually read.
dlt on a 4 GB pod: loading multi-GB datasets without the SIGKILL
Exit code 137 is dlt's way of saying "your RAM and I need to talk"
Migrating Tableau dashboards: Streamlit code vs. Sigma with a parity gate
Same migration, opposite bets: rebuild as Streamlit code you own, or as Sigma workbooks behind a parity gate.
Snowflake Iceberg in 2026: open format, now without the storage chores
Three Iceberg GAs in one week advancing the old tradeoff: open format without babysitting a bucket.
MotherDuck Flights: pipelines an agent can fly, and why dlt belongs in the cockpit
MotherDuck shipped agent-native Python pipelines. The obvious thing to run inside one is dlt.
OpenCode: a terminal coding agent without vendor lock‑in
Bring your own model, ditch the subscription.
How I Stopped ❄️ Cortex Agents from Auto-Picking the Expensive Model
How to configure Cortex Agent models for all agents at once
Infrastructure
Perplexity Search as Code: letting models run the search
Why let a search engine guess what context your model needs? Perplexity's new Search as Code hands the reins to the agent.
Lakehouse
Clippy: one Snowflake Cortex Agent, two faces
The paperclip is back. He works at Pro Juventute, lives in Snowflake, and shows up in both Teams and OpenCode.
Infrastructure
Triggering dltHub Pro pipelines from Snowflake
Moved a dlt pipeline off SPCS onto dltHub Pro. Still want Snowflake to push the button.
Infrastructure
dltHub Pro is here. So is my static-egress proxy VM.
How I kept IP‑locked Snowflake and Azure databases happy with a managed runtime that runs “wherever Python runs”
Infrastructure
On-prem AI agents: the signal in three announcements
Three vendors, one week, one direction: the cloud-only assumption for AI agents is cracking.
Infrastructure
Cortex Guardrails: one boolean, one job, four it doesn’t do
A single boolean on AI_COMPLETE, a Llama Guard 3 wrapper, and a hardcoded filter string. Useful, cheap, and nowhere near enough on its own.
Lakehouse
Cortex Code skills, 90 days in: plugins, not prompts
Cortex Code stopped being “just another AI feature” the moment skills landed. In under three months they’ve become the missing glue: portable, versioned building blocks that turn NPO data stacks into repeatable agent workflows.
Infrastructure
From Orbital GPUs to Snowflake L40S: The Compute Pinch Arrives in the Warehouse
Anthropic is scrambling for GPUs, AMD is finally shipping drop‑in AI cards, and Zyphra just proved you can train real models on a full AMD stack. All of that lands in your Snowflake account the moment you run create compute pool
Lakehouse
Retiring legacy BI with an agentic-built Streamlit app on SPCS
Build-vs-buy got rewritten by agentic coding. So I retired the legacy BI tool. 😎
Lakehouse
Exporting Snowflake data to Excel
Generate XLSX files directly from Snowflake using a small Snowpark stored procedure, writing straight to a stage so stakeholders get ready-to-use Excel workbooks without any CSV exports or manual conversions
Lakehouse
Snowflake interactive warehouses: cheaper, always-warm, and yes, the SQL API works
Snowflake’s interactive warehouses sit in a sweet spot for user-facing workloads: they cost 60% of a same-sized standard warehouse, can stay warm across business hours, and still serve sub‑second lookups from an interactive table
Infrastructure
Two AI agents, one repo: delegating Snowflake work from Claude Code to Cortex Code
Claude Code and Snowflake's Cortex Code turn out to be the same agent wearing different hats - and that's exactly what makes them worth combining. A filesystem handoff between two AI agents replaced hours of manual Snowsight clicking.