Senior Machine Learning Engineer - Perception
Pickle Robot Company builds Physical AI-powered robotic systems for supply-chain truck and container unloading.
Funding history
About Pickle Robot Company
The company provides the Pickle Unload Robot and Dill Autonomy Engine, combining industrial robotics, machine vision, optimal control, generative AI, and targeted human input for warehouse freight handling.
Skills
Candidate Availability
Required and preferred rules are kept separate and reflect the wording in the original posting.
About the Role
You will own end-to-end machine learning work for systems that determine how robots grasp packages. You will formulate modeling problems, build and curate datasets, train and evaluate large vision models, and produce reproducible models ready for production handoff. You will also review code, mentor earlier-career engineers, and contribute to perception and grasping planning.
Requirements
- 5+ years of machine learning and computer vision experience, including taking models from problem statement to measured outcome.
- Experience translating product requirements into engineering tasks and explaining technical trade-offs to non-expert stakeholders.
- Experience building datasets at scale, including generation or collection, curation, automated quality assurance, coverage and bias analysis, and active learning.
- Rigorous evaluation experience with baselines, ablations, and metric design.
- Expert-level Python and PyTorch skills.
- Experience with cloud training infrastructure such as GCP or AWS, Docker, and experiment tracking tools.
- Professional experience developing and training multi-task, multi-modal, or multi-view models, or using self-supervised, outcome-based, or reward-driven learning.
Responsibilities
- Formulate ambiguous picking-performance goals as well-posed machine learning problems.
- Define model predictions, labels, rewards, loss functions, and success criteria.
- Own dataset strategy, including synthetic scene generation, distributed cloud dataset jobs, diversity filtering, and self-supervised label generation.
- Identify and address data bottlenecks.
- Fine-tune large backbones with task-specific heads and run disciplined experiments.
- Define offline and online evaluation metrics for robot performance and edge cases.
- Conduct rigorous code reviews and mentor earlier-career engineers.
- Contribute to the perception and grasping roadmap.
- Produce deployable models with latency budgets, exportable graphs, and reproducible artifacts.
- Partner with runtime integration and inference-optimization engineers for production handoff.
Benefits
- Health insurance
- Dental insurance
- Vision insurance
- Unlimited vacation
- Paid federal and state holidays
- 401(k) contributions of 5% of salary
- Company-covered travel expenses