How We Built an AI Video Editor for Recruiters (Remotion + Claude + Codex)

Sam Nadler:
So I created a little tool that gave our recruiters — who I’m assuming most of them don’t have any video editing experience — the ability to cut up these videos, be completely self sufficient, and generate links to share with our clients.

Jordan Metzner:
Built This Week, breaking it down. Built This Week, we show you how. A fresh idea, a clever tweak, you lock in true. Built This Week.

Sam Nadler:
Hey everyone, and welcome to Built This Week, the podcast where we share what we’re building, how we’re building it, and what it means for the world of AI and startups. I’m Sam Nadler, cofounder here at Ryz Labs, and each and every week I’m joined by my friend, cohost, and business partner, Jordan Metzner. How are you doing today, Jordan?

Jordan Metzner:
Hey Sam. Happy to be back. Another fun, exciting episode. So much going on in the world of AI, and this week is nothing different. New models coming out, exciting news from all the foundation labs, and some updates from NVIDIA as well. It’s been a crazy, exciting week, just like every week has been in the AI race.

Sam Nadler:
Absolutely. Every week there’s something new to talk about, new releases. Before we get into the agenda, don’t forget to like and subscribe. We have new episodes out every single Friday, so stay with us and see what we’re building and what tools are emerging.

This week we’re covering a tool I built for our recruiting team that gives them video editing power so they can be self sufficient and no longer rely on our video editors to cut up candidate videos. I’ll demo the tool, then we’ll talk about how it was built using Remotion, and finally we’ll cover the latest AI news. Does that sound like a plan?

Jordan Metzner:
Awesome. Let’s go.

Sam Nadler:
Alright, quick context. At Ryz Labs, we help source and provide great candidates across engineering, data, product, project management, and more. We interview and evaluate hundreds of thousands of candidates.

When presenting candidates to clients, they typically receive the resume, rate, experience, education — the standard package. But imagine if they could also see the candidate speak for a minute before deciding whether to interview them.

We’ve occasionally created sizzle reels of candidates, but the process was slow. It required our video editors and wasn’t scalable.

So I built a tool that allows our recruiters to generate highlight reels themselves.

It’s powered by our EntreVista AI interview platform. We already have interview videos and transcripts. This tool takes those videos and packages them into a branded sizzle reel.

The recruiter uploads the video, enters the candidate name and title, and pastes in the transcript. The AI analyzes the transcript and automatically identifies the best clips — grouped into categories like English communication, role fit, and personality and drive.

The recruiter can review the suggested clips, trim them slightly, and select the best three. We aim to keep the final reel under one minute — typically around 45 seconds.

Then they preview the branding, choose the aspect ratio, and render the video. Rendering takes about five to ten minutes. That’s dramatically faster than sending it to a video editor and waiting hours or days.

The result is a clean highlight reel that can be shared via an S3 link directly with the client.

We believe this increases the conversion from candidate presentation to client interview. Seeing a candidate speak for 60 seconds changes the dynamic.

This was built primarily in Codex using a tool called Remotion.

Jordan, any feedback?

Jordan Metzner:
It’s awesome. Honestly, it feels like we’re giving away some secret sauce.

Video is a powerful format for evaluating candidates. Historically, editing required heavy desktop tools like Premiere or Final Cut, or mobile apps like CapCut. Now we’re editing video as code.

That’s what makes this interesting. It opens up the possibility of generating infinite video variations programmatically.

Let’s talk about Remotion.

Sam Nadler:
Yeah. You originally told me about Remotion. I started simple, experimenting in Claude Code and Codex, just tweaking visuals with natural language. I have zero video editing background, but the models handled the heavy lifting.

When did you discover Remotion?

Jordan Metzner:
I think of Remotion as a JavaScript video library. Before AI, building video in JavaScript would’ve been painful. Now it’s extremely powerful.

I discovered it during the early vibe coding wave with tools like Lovable and Bolt. It lets you generate dynamic, React-based video compositions.

I’ve used it to create product trailers, narrated slide decks using ElevenLabs for voiceover, and automated documentation videos.

It’s early, but video as code is going to be huge.

Sam Nadler:
Totally agree.

Let’s pivot to the news.

Internally, we’re pushing everyone — technical or not — to ramp up on AI. And this article caught our attention: “You finally figured out AI at work. Now here comes the bill.”

AI output feels free, but behind the scenes it’s data centers and token spend.

Jordan Metzner:
Exactly. We’ve given everyone access to Claude, Claude Code, Claude Cowork, Codex, ChatGPT, Gemini. But not all models cost the same.

We’ve already seen surprise bills. A server left on accidentally cost $35,000. A new Cursor feature led to a $2,000 bill.

Now the question becomes: how do you manage AI usage? How do you measure token input versus employee output?

It introduces a new management layer.

Sam Nadler:
Right. AI usage will be evaluated both in terms of productivity and cost efficiency. Are employees adopting AI responsibly? Are they maximizing impact without being reckless?

We haven’t implemented strict guardrails yet, but we’re probably one surprise bill away from doing so.

Jordan Metzner:
Some companies are already limiting token usage per employee. Others are exploring open source hosting to reduce costs.

At the same time, models are getting cheaper and more powerful. So the efficiency curve keeps improving.

The big question is: what are you getting for your AI spend? Faster roadmap? Fewer engineers? More revenue?

It’s too early to know how it settles.

Sam Nadler:
Totally agree.

Great episode, Jordan. Thanks for reviewing the highlight reel tool and discussing Remotion. Any final thoughts?

Jordan Metzner:
It’s exciting. More people are building at every level. These tools are becoming more accessible and democratized.

Can’t wait to see what next week brings.

Sam Nadler:
See you everyone. Bye.

‍