13 hours ago
Base Salary
$200k - $260k/yr
Responsibilities
- Design and own computer vision architecture for fine-grained item identification.
- Adapt foundation vision models to the marketplace catalog.
- Define accuracy standards and develop calibrated confidence scores that communicate model uncertainty.
- Partner with labeling providers to turn model errors into data-collection priorities.
- Balance expansion into new item categories with accuracy improvements in existing categories.
- Own the path from model research to production serving, including latency, cost, and reliability decisions.
- Explain model capabilities and limitations to technical and non-technical stakeholders.
- Lead technical decisions on an open-ended problem.
Requirements
- At least 5 years of applied computer vision experience, including production systems shipped at meaningful scale.
- Industry experience with image or item identification, object recognition, or detection, including fine-grained or instance-level classification.
- Hands-on production experience fine-tuning and deploying vision transformers or CNNs with PyTorch or TensorFlow.
- Experience adapting foundation vision models, building evaluation frameworks beyond loss metrics, and developing calibrated confidence scoring.
- Experience owning production model-serving decisions involving latency, cost, and reliability.
- Strong communication skills and ability to lead technical decisions.
- Familiarity with low-latency APIs is useful.
- Experience with active learning, human-in-the-loop labeling, visual search, marketplaces, or early-stage startups is a plus.
Benefits
- Hybrid role based in New York, New York, United States.
Tech Stack
PyTorchTensorFlow