Selected Work
Show packages, motion systems, campaign graphics, and visual identities ready for your real credits and imagery.
Selected Work
Apps, prototypes, dashboards, and code projects with space for technical context, links, and outcomes.
graphics / 2024-present
Data-driven broadcast graphics and API extensions
Apex Legends Global Series
Gallery
POI Drop Draft
Gallery
Social Media Graphic
This work was built for Apex Legends Global Series broadcast, where the upstream game data already existed but did not always map cleanly to the kinds of graphics and production tools needed on air.
The opportunity was not just to consume the API, but to extend it into more useful, more readable, and more production-friendly broadcast experiences.
Two of the clearest examples were a point-of-interest drop draft graphic and a custom social media pipeline. Both were designed to turn existing data or content sources into graphics that producers could actually play during a live show.
The POI drop draft system tied into the existing upstream API from the game development team, then extended it with multiple upstream queries to build a richer output. The graphic combined team drop spots, the names of all available locations, map imagery, and team logos into a clearer live draft display, while also supporting cached views of previous drafts for reference and storytelling.
The social media system solved a different production problem. Instead of building manual posts in Figma every time broadcast wanted to feature social content, I built a custom server workflow and Chrome extension that let producers capture Twitter or Reddit posts directly and push them into the graphics pipeline.
That server-side extension flow made it possible to bypass API-key restrictions in the day-to-day production workflow while still keeping post capture fast and simple. The result was a social package that could output both a full-frame graphic and a lower third, depending on what the show needed.
The biggest value of this work was turning existing ALGS data into enhanced broadcast experiences rather than treating the upstream API as a fixed limit.
The POI drop draft graphic gave production a more useful way to explain live strategy and revisit prior draft states, while the social media pipeline made it much faster to get relevant posts to air.
Together, these systems reduced producer friction, replaced slower manual design workflows, and created graphics that were both more responsive to the live show and more directly tied to the underlying game data.
development / 2025
Streamer Bowl VII live ordering and fulfillment
Custom Shop App for Interactive Gameshow
Gallery
Player Shop
Gallery
Control Overview
Gallery
Fulfillment Board
This app was built for Twitch Rivals’ Streamer Bowl VII, a gameshow format featuring top Twitch creators and NFL players.
Players spent a fixed budget on items with limited inventory, while a separate fulfillment team delivered the food, props, or experiences tied to those purchases. Items also unlocked in phases, controlled live by a central show operator.
Before this app existed, the workflow lived in a Google Sheet. It worked, but it was clunky for handlers, difficult to manage cleanly under show pressure, and not well suited to limited inventory across multiple concurrent users.
The system was designed around three connected views: a player-handler shop, an admin control surface, and a fulfillment board.
The shop allowed handlers to place orders against a shared inventory pool, while the backend let production control unlock phases, adjust balances, and manage item availability in real time.
Because multiple users could compete for the same finite items, the purchase flow had to behave transactionally and keep inventory accurate under concurrent use. The fulfillment board then turned those purchases into a clear dispatch workflow, using color coding to help the operations team route deliveries to the correct player groups quickly.
I built the app end to end, including backend logic, real-time updates, persistence, and interface design. It was hosted persistently on my private network behind Cloudflare SSO, and product icons were sourced from artists on my team.
The biggest win was replacing a messy spreadsheet process with a purpose-built live operations tool that was easier for handlers to use and easier for production to control.
That change made ordering and fulfillment more streamlined during the show, while also improving staffing efficiency for the player handler team.
Just as importantly, the app made it practical to run a finite-inventory shop mechanic live without losing trust in the state of the game. Players could spend against real limits, show control could unlock items in phases, and the fulfillment team had a clearer path from order to delivery.
development / 2025-present
Shared GPU server control for live production
EC2 Dashboard
Gallery
Server Overview
Broadcast teams were relying on expensive GPU-backed EC2 servers that often ran around the clock, even when no show was actively using them.
That wasted budget, but it also made ownership messy. Multiple productions could share the same server, and there was no simple internal tool for engineers to see live status, confirm the current IP, or keep billing aligned with the right show.
The dashboard was built as an internal, show-critical control surface for broadcast engineers. It gave the team a reliable way to manage server uptime, track which production was using each machine, and reduce cost without making the workflow harder during live operations.
The product was designed around a single overview screen that engineers could read quickly under production pressure.
From that surface, users could start or stop servers, confirm whether a machine was active, check its current IP when it was not static, and tag shows in or out so shared infrastructure stayed correctly assigned.
The system also updated AWS EC2 tags to keep billing tied to the proper show or project. I handled the full stack build from concept through implementation, then trained the engineers who depended on it in live workflows.
The dashboard turned server control from an ad hoc engineering task into a visible, repeatable workflow.
Broadcast engineers could manage shared EC2 capacity directly, avoid leaving costly machines running unnecessarily, and keep show ownership clearer across overlapping productions.
The biggest measurable result was cost reduction: the tool helped save more than $45,000 per year by making it practical to stop unused GPU-backed servers instead of letting them run 24/7.
development / 2023-present
Stream Acquisition & Graphics Engine
SAGE - Stream Acquision and Graphics Engine
Gallery
Global Overview
Gallery
Distributed Multiview
SAGE was built for remote esports productions where the game still needs to be covered like a traditional broadcast.
Some titles do not offer observer clients or any clean way to view the match outside the players’ perspective. Before SAGE, those shows relied on ad hoc one-feed-per-PC workflows that were difficult to scale, route, and keep visually consistent in a live studio environment.
The system centralizes remote ingest, data aggregation, and output assignment so operators and producers can bring Twitch, YouTube, Kick, and similar player streams into broadcast-ready OBS engines across distributed machines.
Each engine maps to a specific upstream row and can be remapped as production needs change, while graphics updates such as player and team names stay synchronized with the selected output.
The platform was developed incrementally over the last three years across dozens of productions, including World Series of Warzone, and has scaled to 40+ simultaneous outputs feeding a local broadcast switcher.
SAGE was designed as a centralized control plane for distributed OBS execution.
A local app running on each target machine can launch up to eight OBS engines, with every engine mapped to a specific upstream Google Sheet row that defines the active feed and graphics state.
Those mappings can be reassigned live, giving operators a way to redistribute coverage quickly without rebuilding machine-level configurations.
On the ingest side, the system opens supported Twitch, YouTube, Kick, and similar web streams, identifies the source payload, and injects the live stream into OBS browser sources over WebSockets.
From there, SAGE manages the object-level updates needed to treat OBS like a lightweight graphics engine, including swapping large image sets, updating text fields, and changing colors so player and team information stays aligned with the selected output.
Deployment was also part of the design problem, so the platform was built to favor centralized rollout and remote updates rather than one-off studio installs.
The result was a system that replaced fragile one-feed-per-PC workflows with a scalable, repeatable production model for remote esports coverage.
Operators gained a single dashboard for machine status, engine assignment, graphics control, and stream health.
Producers could adapt coverage without rebuilding the show around ad hoc hardware layouts.
Over time, SAGE became the backbone for dozens of productions and proved its value at broadcast scale.
The platform has driven 40+ simultaneous outputs from cloud-managed control to a local switcher, while reducing deployment friction, improving consistency across graphics updates, and making remote games without observer clients practical to cover in a live studio environment.
graphics / 2025
Interactive LED floor games for Twitch Rivals
TwitchCon San Diego 2025
Gallery
Floor Game Overview
Gallery
Ross Control Surface
Gallery
Game States
This project was built for Twitch Rivals at TwitchCon San Diego 2025: Operation Map Breakers featuring Ludwig.
The production wanted to turn a large LED floor into an integral part of the competition rather than a passive stage element. The result needed to deliver visual spectacle for the audience while still functioning as readable, responsive gameplay for players and show staff.
The challenge was technical as much as visual. The floor was not a true touchscreen surface, but the games needed to feel interactive and immediate on air. The production also needed multiple game formats to run on the same footprint without slowing down pacing or requiring a separate hardware system for each mechanic.
We built three custom games for the floor: Codenames, Battleship, and a tile memory game. Ross XPression drove the on-screen experience, while a custom Ross Dashboard gave a competition admin a fast control surface for operating each game live.
My role covered the art direction, dashboard design, and the on-site integration work required to make the system behave reliably across the full 30ft x 30ft LED surface. My team contributed some of the XPression layout work, while I focused on the game interaction model and the overall playable system.
The core trick was using Ross Dashboard to fake a touchscreen floor. Rather than relying on literal touch hardware, the control layer triggered mapped floor states in XPression so the games felt reactive and intentional in real time. That let us use the floor as both a broadcast visual and a competition mechanic.
The project turned an expensive scenic asset into a meaningful part of the show for Twitch Interactive instead of leaving it as background spectacle.
It also gave production three distinct interactive games on one stage system, with a control workflow that was practical to run live under competition pressure.
Most importantly, the floor became integral to the event’s gameplay and identity. The system balanced showmanship with control, making the LED surface feel responsive, deliberate, and central to the competition.