---
title: 'Research Engineer - Data Infrastructure/ML at Third Dimension'
canonical: 'https://feeny.ai/job/research-engineer-data-infrastructure-ml-third-dimension-bay-area-by6r3qxyxjdm'
type: 'job'
last_seen: '2026-09-09'
---

# Research Engineer - Data Infrastructure/ML at Third Dimension

- **Company:** Third Dimension
- **Location:** Bay Area
- **Employment:** full-time
- **Work type:** hybrid
- **Posted:** 2025-10-17
- **Last confirmed live:** 2026-09-09
- **Apply:** https://jobs.gem.com/third-dimension/am9icG9zdDo8foQfIfm9Jv2WHVV6KCdz

## Job description

Third Dimension is building SuperSim, a new kind of simulator which can enable fast, cost-effective and photorealistic 3D simulations directly from source data using AI. We are working with customers across multiple industries, including autonomous vehicles, drones, robots in industrial and manufacturing environments, and more.

## About the Role

We’re looking for a Data Infrastructure / ML Engineer to build the backbone of our 3D AI systems. You’ll design scalable pipelines for 3D, video, image, and other sensor data (e.g., LIDAR) and develop the ML workflows that power SuperSim. Your work will enable our researchers to train, validate, and deploy models faster - ultimately shaping how robots and autonomous systems are tested in the real world. In this role, you will:

- Build and maintain high-performance data pipelines to ingest, transform, and version multi-modal datasets (3D, video, sensor).
- Design and optimize distributed training and data-processing infrastructure - across cloud and containerized environments (Kubernetes, Ray, Dask, EKS, Buildkite).
- Collaborate with researchers to productionize ML models (PyTorch), from prototype to deployment.
- Develop tools and APIs that make data discoverable, reusable, and reproducible.
- Monitor and improve data quality, lineage, and performance across the ML lifecycle.

## Preferred Qualifications & Skills

- Strong programming background in Python.
- Proficiency in distributed data-processing technologies (e.g., Ray, Apache Spark, Flyte, Dask).
- Hands-on experience with cloud infrastructure (AWS, GCP, or Azure), Kubernetes, and distributed training frameworks (Ray, RLLib, PyTorch DDP, or Horovod).
- Knowledge of dataset versioning, experiment tracking, and reproducibility tools (DVC, MLflow, etc.).

Nice to Haves

- Exposure to simulation, robotics, or autonomy testing pipelines.
- Experience with deep learning frameworks and ML workflows (PyTorch).
- Familiarity with 3D or robotics data formats (point clouds, meshes, radiance fields, LIDAR).
- Contributions to open-source ML/infra projects or publications in top-tier AI/ML/CV venues.
- Experience with frontend / UI work.

Location

- London, UK / Graz, Austria/ Bay Area: This is a priority for us, as we're trying to build a presence around our already existing teams in London, Graz or the Bay area. Hybrid (remote + in office days) preferred.
- Remote: If hybrid is not an option, but you're amazing, we're open to having you work out your home location.

## Benefits

- Competitive salary & stock options – Everyone is an owner and shares in our success.
- Pension / retirement plan – Company contributions to support your long-term financial wellbeing.
- Health & wellness – Comprehensive health, dental, and vision insurance (with regional equivalents for employees outside the UK).
- Flexible time off – Generous holiday allowance plus local public holidays.
- Hybrid-first culture – State-of-the-art London workspace with the flexibility to work remotely.
- Workspace support – Latest hardware and home office equipment to set you up for success.
- Learning & development – Budget and time for conferences, courses, and workshops to keep you at the frontier of ML and 3D AI.
- Community & connection – Team offsites, events, and opportunities to connect with global colleagues.

## About Third Dimension

## Company Overview
- **One-liner**: Third Dimension is a spatial generation company that builds high-fidelity 3D worlds to power embodied AI systems, starting with a simulator called SuperSim for robotics.
- **Entity Type**: Private (Seed Stage)
- **Headquarters**: San Francisco, California, United States
- **Founded**: 2024
- **Founders**: Tolga Kart (CEO) and Piotr Sokolski (CTO)

## Core Business
- **Primary industry**: Spatial Generation / 3D Simulation for Embodied AI (Robotics, Autonomous Vehicles)
- **Target customers**: B2B — companies developing autonomous systems, including autonomous vehicles, drones, and industrial/manufacturing robots
- **Mission or purpose statement**: To close the gap between simulation and reality so autonomous systems can operate reliably and safely in the real world.

## Products & Services
- **[SuperSim](https://thirddimension.ai)**: A next-generation simulator that reconstructs the real world in high-fidelity directly from sensor data (LIDAR, video, position data) using neural rendering and Gaussian Splatting. It enables users to debug failures in exact context, re-run software with different behaviors, test rare edge cases, and build thousands of scenario libraries from a single capture — all without a rendering engine or large team of 3D artists. Integrates AI-based world modeling for infinite variations, dynamic environments, and novel object insertion.

## Market Standing
- **Valuation/Market Cap**: Not disclosed
- **Total Funding**: $7M (Seed Round, October 8, 2024)
- **Lead Investor**: Felicis (with 7 investors total)
- **Key Metric**: Total funding of $7M
- **Growth Signals**: Headcount grew 60% year-over-year (from 10 to 16 employees); expanding globally with offices in San Francisco (HQ), London, Istanbul, and Graz (Austria); growing LinkedIn follower base (+223.7% yearly to 1,515 followers); team includes veterans from Tesla, Wayve, Google, Luma AI, NVIDIA, Niantic, and Unity.

## Competitive Advantages
- **Proprietary technology**: Uses advanced 3D techniques like neural rendering and Gaussian Splatting to close the "domain gap" — enabling photorealistic simulations directly from real-world data, unlike traditional rendering-engine-based simulators that look like games and cause performance reliability issues.
- **Full-stack rethink**: Not just a visual improvement; SuperSim is a complete system that allows continuous feeding of real-world data back into the AI training loop, creating a data flywheel.
- **Founding team**: Founders are veterans in 3D reconstruction and AI, with research leadership from Christian Rupprecht (Chief Researcher) and deep technical experience from top AI/robotics companies.

## Strategic Focus
- **Current priority**: Dominating the robotics simulation market by enabling fast, cost-effective, and photorealistic 3D simulations for autonomous vehicles, drones, and industrial robots.
- **Future expansion**: Exploring how spatial generation technology can transform entertainment (gaming and movies) by making 3D content creation faster and more accessible.

## Why Work Here
- **Mission-driven**: Work at the intersection of 3D generation, AI, and robotics — solving a fundamental bottleneck in embodied AI.
- **Culture**: Described as a team of "bold thinkers" with a strong research-first DNA. The team is small (~16 employees) and highly technical, with a significant focus on research (31% of workforce).
- **Offices**: Hybrid/Remote-friendly with physical hubs in San Francisco (HQ), London, Istanbul, and Graz, Austria. Offers both onsite and remote/hybrid options depending on role.
- **Team caliber**: Colleagues come from top AI/robotics companies (Tesla, Wayve, Google, Luma AI, NVIDIA, Niantic, Unity) and top research institutions (University of Oxford). Key executives include a dedicated Chief Science Officer.
- **Open roles**: Currently hiring Research Engineers and Researchers in 3D Reconstruction, Generative AI, SfM/SLAM, and Data Infrastructure/ML across multiple locations.

## Sources
1. [Third Dimension Website](https://thirddimension.ai)
2. [Third Dimension About Page](https://thirddimension.ai/about)
3. [Third Dimension Blog - SuperSim](https://thirddimension.ai/blog/posts/reconstruct-the-real-world-with-supersim)
4. [Third Dimension Careers - Gem](https://jobs.gem.com/third-dimension)
5. [LinkedIn - Third Dimension AI](https://linkedin.com/company/third-dimension-ai)
6. [Built In - Third Dimension AI](https://builtin.com/company/third-dimension-ai)

## Other roles at Third Dimension

- [General Application](https://feeny.ai/job/general-application-third-dimension-bay-area-qsnk4zrpwmp9) — Bay Area
- [Research Engineer - SfM/SLAM](https://feeny.ai/job/research-engineer-sfm-slam-third-dimension-bay-area-m3fy01cg0jje) — Bay Area
- [Researcher - 3D Reconstruction](https://feeny.ai/job/researcher-3d-reconstruction-third-dimension-bay-area-jk5fnvyk8k89) — Bay Area
- [Researcher - Gen. AI](https://feeny.ai/job/researcher-gen-ai-third-dimension-bay-area-fjhtq80rpk06) — Bay Area