About Me

Hi, I'm Ben.

As of 2025 I'm a Principal Engineer at Amazon IVS and Twitch where I've worked since 2019. I focus mostly on AWS IVS's Low Latency HLS product but help out everywhere I can.

Behind the scenes I've built the transcode capacity aggregation and transcode job orchestration systems, and my biggest cross-org effort was leading the multi-year migration of our video infrastructure off of our own on-prem bare metal datacenters/network to AWS. I owned the global capacity roadmap and disaster resilience planning in EC2, and generally lead the infrastructure side of things for IVS.

A couple of user-facing features I've shipped are Timed Metadata and SRT Ingest.

In the past I've worked on Amazon Cloud Cam (RIP) where I worked on our WebRTC SFU and computer vision pipeline. Before that I worked at CENTR Cam (Also RIP) doing things like optimizing image processing kernels to dewarp fisheye cameras and blend them for Movidius' (RIP-ish?) "vector processing units" and generally trying to get things done for my first job out of college.

In college I was interested in GPU programming and programming language design. Still am. Lately I've been interested in learning more about Linux which you can see in my io_uring + Rust async project twoio.

I'd describe myself as a software engineer with solid knowledge in distributed systems, datacenters, and video, who knows how to acquire knowledge when needed and help people get things done. One of my favorite compliments from a manager was "I don't have to worry when I put you on a project because I know you'll move it forward somehow". My most used quote from an old co-worker is "just another day in paradise".

My goal with this blog is to write about things I'm learning about. I will fully admit to not having the deepest knowledge in any one subject, but it's always good to learn and play around with things so you can go deeper when you're required to.

You can contact me at benhirsch24@gmail.com if you'd like.