“Eric is a fantastic engineering lead. His communication skills with team members of all disciplines are excellent. Going beyond just dealing with questions/issues as they occur, Eric will take the time to grow the technical knowledge of the full team to improve efficiency and head off problems before they occur. He goes out of his way to fully understand and prioritize the needs of the end user for whatever systems he is leading or directly implenting. The breadth of his technical expertise is equally impressive.”
About
Activity
-
Almost 1 year has flown by since I joined recruyt , and what a journey it’s been! So much has changed for me, and I can honestly say it’s been one of…
Almost 1 year has flown by since I joined recruyt , and what a journey it’s been! So much has changed for me, and I can honestly say it’s been one of…
Liked by Eric Undersander
-
Most recent checkpoint of n1 vs Opus 4.6! On Navi-Bench and Westworld browser automation benchmarks: - Same accuracy - n1 is 2.5x faster - n1 is…
Most recent checkpoint of n1 vs Opus 4.6! On Navi-Bench and Westworld browser automation benchmarks: - Same accuracy - n1 is 2.5x faster - n1 is…
Liked by Eric Undersander
Experience & Education
Publications
-
Block-Sparse Recurrent Neural Networks
Recurrent Neural Networks (RNNs) are used in state-of-the-art models in domains such as speech recognition, machine translation, and language modelling. Sparsity is a technique to reduce compute and memory requirements of deep learning models. Sparse RNNs are easier to deploy on devices and high-end server processors. Even though sparse operations need less compute and memory relative to their dense counterparts, the speed-up observed by using sparse operations is less than expected on…
Recurrent Neural Networks (RNNs) are used in state-of-the-art models in domains such as speech recognition, machine translation, and language modelling. Sparsity is a technique to reduce compute and memory requirements of deep learning models. Sparse RNNs are easier to deploy on devices and high-end server processors. Even though sparse operations need less compute and memory relative to their dense counterparts, the speed-up observed by using sparse operations is less than expected on different hardware platforms. In order to address this issue, we investigate two different approaches to induce block sparsity in RNNs: pruning blocks of weights in a layer and using group lasso regularization to create blocks of weights with zeros. Using these techniques, we demonstrate that we can create block-sparse RNNs with sparsity ranging from 80% to 90% with small loss in accuracy. This allows us to reduce the model size by roughly 10x. Additionally, we can prune a larger dense network to recover this loss in accuracy while maintaining high block sparsity and reducing the overall parameter count. Our technique works with a variety of block sizes up to 32x32. Block-sparse RNNs eliminate overheads related to data storage and irregular memory accesses while increasing hardware efficiency compared to unstructured sparsity.
Other authorsSee publication -
The Sticky Collision Beam
Game Developer Magazine
See publicationCreating intelligent third-person cameras is rife with difficulty. There are enemies to keep in view, obstacles to avoid, and important objects to highlight. Here, engine programmer Eric Undersander proposes a camera with a sticky collision beam to bypass many of the major issues faced by AI-controlled cameras.
Recommendations received
9 people have recommended Eric
Join now to viewMore activity by Eric
-
As we enter the new year of 2026, I’d like to share a brief career update. I have moved on from FAIR/Meta after an incredible 3.5 years. FAIR was my…
As we enter the new year of 2026, I’d like to share a brief career update. I have moved on from FAIR/Meta after an incredible 3.5 years. FAIR was my…
Liked by Eric Undersander
-
https://lnkd.in/gYzJiCgx It doesn’t replace hard work or vision, it lets you accelerate it.
https://lnkd.in/gYzJiCgx It doesn’t replace hard work or vision, it lets you accelerate it.
Liked by Eric Undersander
-
I’m happy to share that I’m starting a new position as Lead Animator at Big Huge Games / Xsolla! Thank you to all the people who have helped me…
I’m happy to share that I’m starting a new position as Lead Animator at Big Huge Games / Xsolla! Thank you to all the people who have helped me…
Liked by Eric Undersander
-
Sometimes I wonder what happened to all the companies that were planning to automate self-driving trucking. So I looked into it and wrote a bit about…
Sometimes I wonder what happened to all the companies that were planning to automate self-driving trucking. So I looked into it and wrote a bit about…
Liked by Eric Undersander
-
I had an incredible time in Budapest with our global teams at the StoneX Payments Offsite. It was my absolute pleasure to present on our US and…
I had an incredible time in Budapest with our global teams at the StoneX Payments Offsite. It was my absolute pleasure to present on our US and…
Liked by Eric Undersander
-
I'm hiring an intern to join our US Payments compliance practice in Miami over the summer! https://lnkd.in/ecGsxCAx
I'm hiring an intern to join our US Payments compliance practice in Miami over the summer! https://lnkd.in/ecGsxCAx
Liked by Eric Undersander
-
Today we’re releasing the Habitat-Matterport 3D Semantics dataset, the largest public dataset of real-world 3D spaces with dense semantic…
Today we’re releasing the Habitat-Matterport 3D Semantics dataset, the largest public dataset of real-world 3D spaces with dense semantic…
Liked by Eric Undersander
-
After 4 years of work, I am ecstatic to announce that my new book “Artificial Intelligence in Schools” has been picked up by probably the best…
After 4 years of work, I am ecstatic to announce that my new book “Artificial Intelligence in Schools” has been picked up by probably the best…
Liked by Eric Undersander
Other similar profiles
Explore top content on LinkedIn
Find curated posts and insights for relevant topics all in one place.
View top content