Waabi, founded by AI pioneer and visionary Raquel Urtasun, is an AI company building the next generation of self-driving technology. With a world class team and an innovative approach that unleashes …
📋 Job Overview
Waabi, founded by AI visionary Raquel Urtasun, is the leader in Physical AI. With a world-class team, we're unlocking the next era of autonomous transportation with technology that's powering commercial autonomous trucks and robotaxis. Waabi is backed by and partners with world leaders in AI, automotive, logistics, and deep tech.
🏢 About Waabi
With offices in Toronto, San Francisco, Dallas, and Pittsburgh, Waabi is growing quickly and looking for diverse, innovative and collaborative candidates who want to impact the world in a positive way. To learn more visit: www.waabi.ai
🎯 The Role
As TLM of the ML Datasets Platform, you will play a critical role defining the APIs, tooling, and storage formats that make datasets a shared, first-class asset at Waabi: versioned, efficiently updatable, fully traceable, and fast to load across GPU workloads. The technical challenge is to find the right abstractions so that every ML team can integrate with the platform for its model development use case without being forced into a one-size-fits-all mold, and without building its own data infrastructure on the side. You will do this as a hands-on technical leader who also grows and manages the team, sets its roadmap, and partners closely with the ML Training Platform team and the ML development teams who are the platform's customers.
✅ Key Responsibilities
Lead and grow the ML Dataset Platform team as both its technical lead and people manager: set the technical direction, define the roadmap, hire, and mentor engineers.
Define the APIs, tools and storage formats for Waabi's ML datasets, balancing maximal reuse of data and tooling against the specific needs of each training workload.
Build the tooling to manage datasets at scale, including versioning and efficient incremental updates without full rebuilds.
Design the APIs that let teams express training-time transformations along the continuum between fully reproducible ahead-of-time ETL and dynamic transformation within data loaders.
Own the common dataloading tools used across workloads, including observability into bottlenecks and the throughput of training jobs.
Streamline data access for multi-cloud training so that datasets are available as needed with predictable cost and latency.
Make data lineage a first-class, fully automatic property of every dataset, from raw data through to the model trained on them.
Partner with the ML Training Platform team on the boundary between data and training infrastructure, and with the ML development teams on transitioning shared-dataset tooling into the new platform.
Drive adoption: work directly with model teams to migrate their training data onto the platform and feed what you learn back into your team’s roadmap.
📌 Required Qualifications
Deep ML dataset experience: You have designed and built large-scale data infrastructure for ML training (storage formats, dataset versioning, ETL, dataloading) and understand the performance characteristics of petabyte-scale sensor or multimodal data.
