Data · Engineering · Product Design

Newsroom
Tools

I work inside live newsroom systems — adding new data pipelines and templates, prototyping better workflows, and shipping lightweight features on top of existing tools — so journalists can turn data into stories under real broadcast pressure.

How I build

I drop into established codebases, find what's slowing people down, and build on top of them — using tools like Cursor and AI-assisted coding to quickly understand complex systems, spot patterns, prototype ideas, and build lightweight features, then working with engineers to test and release them reliably.

These are internal CNN production tools and are not shown publicly. The diagrams below are simplified illustrations of the workflows — with the parts I contribute to highlighted.

01 — Data Pipelines & Templates

Element Builder

A producer-facing tool for turning live data into consistent on-air graphics. I extend it with new data sources and templates.

Data sources
Template selection
Docs & guides
Producer inputs
On-air graphic

Where I plug in: new data sources (built as Airflow DAG pipelines), new interactive graphic templates, and how-to docs & guides.

The problem

Live coverage constantly needs new datasets and new graphic formats. The tool has to keep absorbing fresh data feeds and templates in a breaking news environment — fast, and without breaking on air.

What I did

I bring new data sources and APIs into the tool as Airflow DAG pipelines, and build new graphic templates and one-off special projects on top of the existing system — so producers have the datapoints and formats they need for whatever story is breaking.

How I work

I pick up the existing codebase and pipeline setup quickly and build within it — pairing with engineers to ship new data feeds and templates reliably, and extending the tool as coverage needs evolve.

Docs & training

Wrote how-to guides and ran training sessions with teammates so producers could confidently use the new templates and tools.

Impact

Kept the tool current with the stories being told, giving producers a faster, more consistent way to turn data into graphics on air.

Svelte JavaScript REST APIs Airflow Python Git

02 — Product Design & Prototyping

Producer Tools 2.0

A redesign of the content system producers use to manage CNN's Magic Wall Election App — from clunky to intuitive.

Editorial input
CMS
Election app
Broadcast

My focus: redesigning the CMS interface and workflow — the editorial-input step where producers actually work.

The problem

Producer Tools 1.0 — the system editorial teams use to manage on-air election graphics — wasn't intuitive and was hard to use under deadline, slowing people down exactly when accuracy and speed matter most.

What I did

With the team, we reimagined it as Producer Tools 2.0 — prototyping the redesign in Figma and rethinking both the look and the functionality so the workflow made sense to the people using it live.

How I work

I focused on the experience — identifying the usability problems, designing and prototyping the solution, then partnering with engineers through build and rigorous QA to get it right.

Documentation

Wrote how-to guides documenting the redesigned workflow and how it works.

Impact

A clearer, more intuitive content tool that made managing on-air election graphics faster and less error-prone.

Figma Prototyping UX Design QA Jira Confluence

03 — Data, Features & QA

Magic Wall Election App

The flagship CNN Magic Wall Election App anchors use on election nights — I feed it data, build lightweight features, and hammer on the QA.

Anchor requests
Data & analysis
Rehearsals & QA
Live data → app
On-air interaction

Where I plug in: election data & analysis, rehearsals & QA, and lightweight in-app features.

The problem

On election night, anchors need to explore results live — vote margins, shifts, demographics — instantly and accurately, on air, in front of millions of live viewers with no room for error.

What I did

I manage election and demographics data in the app — voting data, win-margin analysis, and topical datasets spanning cost of living, health, and demographics — run on-the-fly analysis during live coverage, and develop lightweight new features built around that data.

How I work

I work on top of the existing app — learning the codebase and extending it with data, analysis, and focused features, backed by rigorous QA before and during air.

Impact

Helped CNN viewers understand election results in real time through accurate, clear, interactive graphics.

JavaScript Python SQL Data Analysis QA D3 Git