Case study · 2026

One role across revenue, product, and the live event.

I joined RenderATL as its only technical revenue hire. I worked across sales, marketing, and engineering to source customers, close deals, and build the platform they bought.

Context

The business needed sponsor revenue and the software required to deliver what sponsors and attendees expected. The work crossed departments because the handoff between a signed agreement, sponsor activation, event access, and post-event follow-up was part of the product.

My role

I carried the customer conversation from prospecting through delivery. That meant selling, translating sponsor and operations needs into a build plan, shipping the software, and supporting it under live event load.

What shipped

  • A six-application platform spanning analytics, sponsor workflow, authenticated streaming, annotation, people operations, and gamification.
  • A production follow-up agent for seven representatives. It gathers CRM, call-note, and email context, drafts in the representative's voice, and cannot send.
  • The gated streaming experience with ticket entitlements and authenticated playback.
  • Session RSVP with capacity limits and waitlisting across the conference schedule.
  • An Airtable-to-Postgres migration across 11 first-tier entities with reconciliation gates, lineage, schema-drift checks, and scheduled validation.
  • RenderATL Rush, a browser game and brand campaign that connected gameplay and sharing to single-use ticket codes.
  • The full curriculum and three workshops for RenderATL University.

Results

$224,500contracted revenue across 11 accounts in 15 weeks
$7,012.50ticket revenue attributed to the RenderATL Rush game
29.7%Rush code-redemption rate
164registrations across the RenderATL University series

Engineering decisions

Human approval stayed in the outbound-sales loop. Model calls had batch-level cost caps, and policy checks blocked off-brand language. The game issued discount codes outside the score-submit transaction so leaderboard contention did not lock inventory work. The database cutover stopped on dirty data instead of treating reconciliation as a post-launch cleanup task.

Related work

Signal Reel is separate. I built the local-first video agent as an independent project. It analyzes transcripts and frames, validates edit decisions, and renders video with FFmpeg. It belongs in the same agent-systems portfolio, but I do not present it as RenderATL client work.

Evidence boundary

The commercial metrics and internal system descriptions are owner-verified employment records. Public links are included only where the destination is still available. Customer records, private source code, contracts, and internal operating data are not published here.