Saša Tomić

About

I am a research engineer and technical lead with 15+ years in distributed systems and storage, across academic research and production. I now work as an independent consultant; what that means in practice is on the services page.

I’m Saša Tomić (a.k.a. Sasha Tomich or Sascha Tomic or Sasa Tomic, depending on region and spelling).

I grew up in Serbia, graduated and got my Dipl. Eng. at the Serbia’s top Belgrade University, School of Electrical Engineering, Computer Science and Informatics Department.

After getting my Dipl. Eng. (equivalent to an M.Sc.), I went to Barcelona to further dig into how computers work, and I specialized with a PhD in computer architecture, where I joined a team led by Adrián Cristal and Osman Sabri Ünsal (under the famous Mateo Valero), researching Hardware Transactional Memory in a newly formed joint research institute between Barcelona Supercomputing Center (BSC) and Microsoft. It was fun and I learned a lot. I completed teh PhD at the BSC and Universitat Politècnica de Catalunya (UPC) with Summa Cum Laude, made a few solid publications at top-tier confierences for the field (MICRO, ASPLOS), which later got me a position at IBM Research in Switzerland, as PostDoc. Two years later, I got a permanent position as a Research Staff Member (RSM). I worked on the firmware for the IBM FlashCore Module, which I enjoyed a lot. I also got an IBM Corporate Award among other smaller awards for the work there (our work got IBM a few $Billion revenue). In the last ~2 years at IBM I worked on Storage Insights, adding streaming ingestion of FlashSystems performance metrics for Proactive Analytics, or AIOps as some call it. The best part of the work was the people. Amazing, brilliant bunch.

At the end of 2019 DFINITY offered me a position to build the next-generation cloud that runs on a decentralized network of machines. That sounded challenging, and my knowledge of storage could help them. I joined them, worked on various layers of the ICP stack for ~2 years before the launch (this includes core execution, IC-OS VMs and reproducible builds, CI and infra), and then we launched the Mainnet. I took the steering wheel for the Mainnet Reliability team, which I led until mid-2025. At this point DFINITY leadership realized that AI is “eating the world” and started building Caffeine.ai as a no-code platform for building ICP applications. For this they needed a LOT of storage which ICP itself couldn’t provide, and they assigned me to build the storage stack that would cleanly integrate into the ICP and give ICP apps a safe cryptographically protected disaggregated storage. It was a worthy task, and we had a production-ready prototype in roughly 6 months when the product went live. After 6 months more the Immutable Blob Storage for ICP was fully operational for all ICP applications, beyond Caffeine.ai. My job there was done, and I was finally ready to look for new challenges.

These days I:

You can find me on LinkedIn, GitHub, or Google Scholar.