Aleksa Gordić

Aleksa Gordić researches large-language-model pretraining and scaling laws on NVIDIA’s Nemotron team. He previously worked at DeepMind and Microsoft and shares practical deep-learning research and engineering tutorials.

Career

Cosigned byLoading...

CosignedSign in to view

Based in: San Francisco
Elsewhere

NVIDIA

Aug 2026 — Present

Research Scientist / Tech Lead, Nemotron Pretraining

Essential AI

Sep 2025 — Aug 2026

Research Scientist / Tech Lead

P-1 AI

Sep 2024 — Jul 2025

Co-Founder and Head of AI

Google DeepMind

Dec 2021 — Mar 2023

Research Engineer

Microsoft

Software Engineer II (Machine Learning)
Sep 2020 — Jul 2021
Belgrade, Central Serbia, Serbia
Worked on HoloLens 2 and related eye-tracking and eyeglass-detection technology.
Software Engineer (Machine Learning)
Sep 2018 — Sep 2020
Belgrade, Central Serbia, Serbia

Telocate GmbH

Jul 2017 — Sep 2017

Software Engineer

University of Belgrade, School of Electrical Engineering

Electrical Engineering, computer science and electronics focus

University of Belgrade, School of Electrical Engineering

Master's studies in electrical

Products & media

aleksagordic.com

Inside TPU and GPU Clusters

An illustrated explanation of how cluster topology shapes collective communication during distributed model training and inference.

Writing

Profile name

Company name

Company name

Profile name

Company name

Company name

Profile name

Company name

Company name

Profile name

Company name

Company name

Profile name

Company name

Company name

Aleksa Gordić

NVIDIA

Essential AI

P-1 AI

Google DeepMind

Microsoft

Software Engineer II (Machine Learning)

Software Engineer (Machine Learning)

Telocate GmbH

University of Belgrade, School of Electrical Engineering

University of Belgrade, School of Electrical Engineering

Products & media

Inside TPU and GPU Clusters

Curated Lists

List Inclusions