Context
Webedia produces and distributes massive volumes of video content for social networks every day. An increasing share of value relies on the ability to quickly cut long films and programs into short clips, adapted to platform usage (YouTube, Facebook, Snapchat, TikTok…).
Historically, this clip production was highly manual, dependent on the expertise of a few senior editors, with high quality variability and difficult deadlines to meet at scale.
Challenges
- Non-scalable clip production facing increasing volumes
- Strong dependence on key profiles (non-formalized expertise)
- Heterogeneous processes across teams and projects
- Difficulty industrializing without degrading editorial quality
- Very strict validation: a single invalid clip renders a batch unusable
Intervention
Design and management of the EdiThor project, an internal software aimed at automating video clip generation from long-form content.
My role covered:
- Functional framing of the product (objectives, editorial constraints, validation criteria)
- Technical architecture design
- Automation of cutting, processing, and video export steps
- Orchestration of production and validation workflows
- Integration with existing tracking and management tools (databases, automations, reporting)
The approach was deliberately pragmatic: automate as much as possible while leaving human control where it’s critical.
System Implemented
- Automated pipeline for generating clips from films
- Fine-tuned cutting rules (durations, formats, platform constraints)
- Validation system for quickly identifying non-compliant clips
- Integration with management databases for batch and iteration tracking
- Continuous improvement loops based on field feedback
Results
- Ability to produce clip volumes far exceeding manual processes
- Reduced dependence on a few key experts
- Progressive standardization of quality criteria
- Significant time savings for operational teams
- Technical foundation ready for larger-scale industrialization
The project is technically operational, with adjustments still needed to reach a reliability level compatible with “zero defect” validation.
What This Project Illustrates
- My ability to work on complex production systems, not just tooling
- An approach oriented toward scalability, robustness, and field reality
- Managing projects where automation must coexist with strong editorial requirements
- The CTO / Head of Systems role on business-critical topics
- My ability to intelligently integrate useful AI tools into business processes
Next Steps
EdiThor is a strategic building block for industrializing short video content production. The natural next steps involve improving reliability, reducing required passes, and extending the system to new formats and platforms.