AI / Agentic Systems
Content Engine: Building a Brand-Aware AI Content Production System
Content Engine is an internal content-production system built around a simple idea: creative reasoning can be flexible, but the production pipeline should be structured, resumable, inspectable, and able to fail safely.
The problem
AI can generate endless content. That is not the same thing as having a reliable production system. Real client work needs brand context, approved media, repeatable formats, quality control, predictable outputs, and a way to stop when something is wrong.
My role
I designed and built the production workflow, from structured briefs and brand context through asset resolution, orchestration, rendering, output audits, and review. Performance and correction feedback are part of the design; the complete learning loop remains in progress.
Production architecture
The production path uses a file-based Node.js orchestrator and structured Brief.v1 JSON. Real source media comes first. Asset generation fills specific gaps, and the selected format is rendered with Remotion. Validation, output audits, and human review are explicit parts of the workflow.
- Brand context
- Topic / formula
- Structured brief
- Validation
- Asset resolution
- Missing-asset generation
- Render
- QA / audit
- Human review
- Delivery
- Feedback
What I built and designed
- File-based Node.js orchestration with budget and wall-time limits
- Remotion rendering and resumable Lambda batch production
- Strict structured Brief.v1 JSON contracts
- Schema validation
- Brand/context loading
- Real-footage-first asset resolution
- Missing-asset generation only when needed
- Resumable stages with service-specific concurrency limits
- Per-brand storage and delivery adapters
- Render and output QA / audits
- Human review
- Performance/correction feedback design
Quality gates that can stop production
The workflow is designed to stop or route for review when the input or output is not usable. A generated file alone is not enough to call a job complete.
- Invalid brief or schema mismatch
- Missing files or assets
- Unusable media
- Bad duration or cue alignment
- Broken render
- Severe visual QA failure
Resumable production and delivery
A failed upload should not require paying to generate every asset again. The orchestrator can resume from individual stages, limits spend and elapsed time, and applies concurrency limits for each service. Remotion Lambda batch rendering also supports resuming completed work.
The rendering path supports per-brand storage, including Cloudflare R2 and Google Drive delivery. The repository documents batch ad production for Luxe’s, connecting the production tooling to actual creative work.
Implementation status
The brief, asset, rendering, audit, and delivery paths are implemented, including documented batch ad production. Feedback and rule-proposal components exist, while the complete integrated production-and-learning loop remains in progress.
Tools and working methods
- Node.js
- Remotion / Remotion Lambda
- FFmpeg
- Cloudflare R2 / Google Drive
- Brief.v1 JSON
- Schema validation
- Asset orchestration
- Human review
Result
Working production components for structured, resumable content workflows, with rendering, validation, audits, and human review. The full pipeline should not be read as a completed autonomous production-and-learning loop.
What I learned
I approach AI production as a systems problem: contracts, inputs, state, validation, reproducibility, quality, and human judgment — not just prompts.