PeopleofSuperintelligence

Researcher

Alexander Toshev

Also known as A. Toshev, Alexander T Toshev, Toshev Alexander

Alexander Toshev is an AI researcher whose work spans computer vision, image captioning, robotics and multimodal systems. He coauthored DeepPose and Show and Tell, helped lead Google Robotics navigation and SayCan, and coauthored Apple's research on multimodal LLMs and data filtering. In 2025, he coauthored UINavBench, a framework for evaluating interactive digital agents. As of October 2026, he is Research Director at Wayve, building foundation models for general robotics, robot manipulation and mobility.

Now at
Wayve
Works
279
Citations
19,831
h-index
34
Badges
Links
Google Scholar

Publication numbers from OpenAlex, as of 22 October 2026.

Timeline

  1. 2014

    DeepPose for human poses

    In 2014, he coauthors DeepPose, which formulates human-pose estimation as regression over body joints. A cascade of deep neural networks improves the precision of its pose estimates.

  2. 2015

    Show and Tell captions

    In 2015, he coauthors Show and Tell, a generative model with a deep recurrent architecture that combines computer vision and machine translation to generate natural sentences describing images.

  3. 2017–2021

    Google Robotics navigation

    From 2017 to 2021, he initiates and co-leads Google Robotics' navigation effort, which produces work on object-driven and social navigation and mobile manipulation. In 2021, he and Anthony Francis describe Google's iGibson Challenge, a real-world setup for testing navigation policies trained in photo-realistic simulation environments.

  4. 2022

    SayCan grounds language in robotics

    In 2022, he co-designs and co-leads Google Robotics' SayCan effort. The paper grounds natural-language instructions in pretrained robot behaviors so the proposed actions fit a robot's physical context.

  5. 2024

    Apple's MM1 multimodal models

    In 2024, he coauthors Apple's MM1 paper on multimodal LLM pretraining. MM1 combines dense and mixture-of-experts models with up to 30 billion parameters, and its authors study architecture and training-data choices.

  6. 2024

    Data Filtering Networks

    In 2024, he coauthors Data Filtering Networks, a study of learned filters for large, uncurated image-text data. The work introduces DFN-5B and uses it to train CLIP models.

  7. 2025

    UINavBench for digital agents

    In 2025, he coauthors UINavBench, an ICCV paper that provides a framework to evaluate interactive digital agents.

  8. As of October 2026

    Wayve's robotics models

    As of October 2026, he is Research Director at Wayve, where he builds foundation models for general robotics beyond vehicles, including robot manipulation and mobility.

Career

  1. NowWayveResearch Director
  2. 2022–2025Apple
  3. 2017–2022GoogleCo-lead, navigation, Co-lead, SayCan
  4. 2006–2010University of Pennsylvania

From the biography and OpenAlex affiliations (the years of papers that list the organization), so some years are approximate.

  1. Google alumni have founded or led 133 other organizations, more than any other AI labIncludes Alexander Toshev. We estimate that alumni of 17 major AI labs went on to found or lead 350 other organizations, including 226 they founded, and we trace 472 career moves directly between 23 labs since 2010.
  2. Of 943 people shaping AI who worked at Google, 446 work there nowIncludes Alexander Toshev. Our estimates of where the 943 people shaping AI who worked at Google and Google DeepMind work now: 446 at Google and 452 across 278 other organizations.
  3. Google had the most people shaping AI every year from 2014 to 2025Includes Google. Our estimates of which organizations employed the most people shaping AI each year from 1955 to 2025, tracing the shift from university labs to companies.
  4. Google has the highest combined h-index among AI organizationsIncludes Google. We estimate that Google holds the highest combined h-index at 8,016, while universities account for 73% of the deduplicated total across AI organizations.

Colleagues

People with dated positions at the same organization in the same years.

Sources and credits

Photo: We have no published photo of Alexander Toshev yet. Send us a free-licensed photo.

  1. DeepPose: Human Pose Estimation via Deep Neural Networks openaccess.thecvf.com. Checked 20 October 2026.
  2. Show and tell: A neural image caption generator research.google. Checked 20 October 2026.
  3. Home sites.google.com. Checked 20 October 2026.
  4. Presenting the iGibson Challenge on Interactive and Social Navigation research.google. Checked 20 October 2026.
  5. SayCan: Grounding Language in Robotic Affordances say-can.github.io. Checked 20 October 2026.
  6. MM1: Methods, Analysis & Insights from Multimodal LLM Pre-training machinelearning.apple.com. Checked 20 October 2026.
  7. Data Filtering Networks machinelearning.apple.com. Checked 20 October 2026.
  8. International Conference on Computer Vision (ICCV) 2025 machinelearning.apple.com. Checked 20 October 2026.