AI workflow
Two ways I bring AI into production: a live agent demo, and the framework I use to roll AI into a team.
I design and orchestrate AI agents that hold a real conversation with a producer, draft client-ready email, and proactively surface the outstanding threads that still need an answer.
On a live production, that means the agent watches the inbox for me, tells me what is waiting, takes an instruction in plain language, and writes the reply in my voice. Below is a walkthrough of how that plays out: a producer checking in, the agent flagging what is outstanding, and replies going out in seconds.
Hi Masharzi, welcome back.
Here's your morning check-in.
Simulated demo: no live API calls, no real emails sent.
The above is a simulated walkthrough. Below is real client work, password protected.
Creative production
Built a custom After Effects workflow for Meta's Video Production team on a Ray-Ban commercial. Created 3D stereoscopic backgrounds in Blender, then wrote an AI-assisted script that auto-loaded every setting the moment the file hit After Effects, and integrated the green screen elements into the final composited output.
A 3-step walk-through of the framework I actually use.
Step 1. Pick a workflow pain point