Deep Learning Engineer - Manipulation

About the Role

In this role, you will work on all aspects of training capable manipulation policies, be it pre-training of a base policy on a diverse multi-embodiment corpus of manipulation trajectories, fine-tuning models to perform a specific task well, curating data collection processes or exploring productive ways to use synthetic data.

This is primarily a deep learning-focused role, so we are looking for experience solving real problems using modern neural networks, and experience in robotics isn't strictly required. However, if you don't have such experience, be prepared that you'd need to familiarize yourself with a new domain quickly.

What You'll Do

• Train policies via representation learning, behaviour cloning and RL; own the full loop from data to deployment.
• Partner with teleoperations to drive data collection: specify what 'good' looks like, ensure diversity/coverage, and close the gap between sim and real.
• Run pre-/mid-/post-training on multimodal LLM/VLM/VLA stacks; plug in new modalities (vision, audio, proprioception, LiDAR/point clouds, …) without breaking existing ones.
• Build and maintain continuous pipelines: ingest simulation + tele-op logs, version them, apply weak-supervision labelling, curate balanced datasets, and auto-surface fresh failure cases into retraining.
• Work with MLOps & Data Platform teams to scale distributed training and optimize models for real-time edge inference.

We're Looking For

• 3+ years building deep-learning systems (industry or research) with shipped models or published artifacts to show for it.
• Hands-on with at least one of: LLMs, VLMs, or image/video generative models - architecture, training, and inference.
• Experience with deep learning infrastructure: streaming datasets, checkpointing & state management, distributed training strategies.
• Strong Python + PyTorch/JAX; you can profile, debug numerics, and write maintainable research code.
• You document experiments clearly and communicate trade-offs crisply.

Nice to have

• Robotics or autonomous driving experience.
• RL for LLMs or robotics (PPO, DPO, SAC, etc.).
• Proven productization of deep nets (latency/throughput constraints, telemetry, on-device optimization).
• Publications at ICLR/ICML/NeurIPS or equivalent open-source contributions.
• Familiarity with OpenVLA, Physical Intelligence (π) models, or similar open VLA frameworks.

What We Offer:

• Competitive salary plus participation in our Stock Option Plan.
• Paid vacation with adjustments based on your location to comply with local labour laws.
• Travel opportunities to our Vancouver and Boston offices.
• Office perks: free breakfasts, lunches, snacks, and regular team events.
• Freedom to influence the product and own key initiatives.
• Collaboration with top-tier engineers, researchers, and product experts in AI and robotics.
• Startup culture prioritising speed, transparency, and minimal bureaucracy.

How to Apply:

For more information on the role, or an informal discussion regarding opportunities we have available, please contact Alicja Szymanska.

Why work with Proactive?

Proactive Global is an industry leading, specialist engineering recruitment agency focused on the automation, manufacturing and advanced technology sectors. We offer specialist recruitment services to a niche customer base, vetting that our clients offer the best opportunities for your career.

Proactive encourages and promotes equality and diversity within the workforce. We act with honesty, integrity and impartiality, ensuring your application is considered on its own merits and without bias.

Proactive Global is committed to equality in the workplace and is an equal opportunity employer.