← ALL WORK
AI & media

VideoForge

From a creative brief to a media pipeline.

AI video generation and media processing with asynchronous rendering and publishing integrations.

Next.jsTypeScriptPythonFastAPIPostgreSQLFFmpeg
MY ROLE

Product & Full-Stack Engineering

BUILT FOR

Creators and teams assembling repeatable video workflows

01 / THE CONTEXT

From a creative brief to a media pipeline.

Video generation is a multi-step process involving models, rendering, storage and publishing. These long-running operations need a reliable workflow rather than one large synchronous request.

What I built

I connected a Next.js frontend to FastAPI services, Celery jobs and FFmpeg processing. Model access through LiteLLM and OAuth-based publishing integrations bring generation and delivery into the same workflow.

02 / PRODUCT CAPABILITIES
01

Generation orchestration

AI-assisted generation workflows coordinate model requests with subsequent media-processing steps.

02

Asynchronous media work

Celery and FFmpeg handle rendering and processing outside the interactive web request.

03

Publishing connections

OAuth integrations with YouTube and Google Drive connect content creation to distribution and storage workflows.

Under the surface.

Follow a real product workflow through its application boundaries. Select any step to explore the implementation.

ENGINEERING IN MOTION

A creative brief becomes a staged media job.

A Celery chain coordinates script generation, speech, composition and upload, with a separate publishing integration.

STEP 01 / CREATIVE INPUT

Project brief

A project defines the material to create. Existing scripts or uploaded voiceovers can be reused by later pipeline stages.

WHY IT MATTERS

Stable stage boundaries allow reused inputs, long-running work and meaningful failure states. Rendering, storage and publishing have distinct responsibilities.

Conceptual flow based on the implementation. Animated connections illustrate data movement, not live traffic.

04 / ENGINEERING JUDGMENT

Built around the constraints.

Model video creation as a pipeline. Isolate expensive rendering from the interface and make external integrations explicit boundaries.

Discuss the engineering ↗

Keep exploring.

Next project ↗