Machine Learning Engineer

Chooch•San Mateo, CA

19d•Hybrid

About The Position

Chooch is seeking a US-based Machine Learning Engineer for a full-time, hybrid role involving frequent hands-on work with customer projects in the Bay Area. The engineer will design, train, configure, deploy, and refine computer vision models and applications for customers in healthcare, manufacturing, and retail. The role includes customer interaction, collaboration with solutions engineering and product teams, and opportunities to lead technical projects and improve processes. The position focuses on developing computer vision systems from data preparation through production deployment and ongoing monitoring. The engineer will adapt models to customer use cases, maintain end-to-end data pipelines, optimize models for production and edge hardware, build evaluation tools, and use production results to identify drift and guide retraining. The role also involves translating customer needs into measurable technical requirements and contributing clean, tested code and shared machine learning tools. Chooch AI is a computer vision platform that uses fixed cameras and AI to monitor hospital supply rooms and inventory. It tracks supplies and updates counts in real time, generates replenishment orders when stock falls below specified thresholds, and helps reduce stockouts, optimize inventory, and improve staff productivity.

Requirements

3–5 years of industry experience building machine learning systems, including at least 2 years focused on computer vision.
BS or MS in Computer Science, Electrical Engineering, Math, or a related field, or equivalent practical experience.
Strong Python skills and hands-on experience with PyTorch or TensorFlow for training and debugging deep learning models.
Experience shipping at least one computer vision model to production and explaining the trade-offs involved.
Solid understanding of CNN and transformer architectures for detection, segmentation, and classification.
Experience with OpenCV and image/video processing, including stream decoding, preprocessing, and augmentation.
Working knowledge of inference optimization, such as ONNX, TensorRT, quantization, or similar methods.
Comfort with Docker, Git, Linux, and GPU training in a cloud environment such as AWS or Azure.
Sound machine learning fundamentals, including loss functions, overfitting, class imbalance, evaluation metrics, and experiment design.
Clear written and verbal communication skills, including the ability to explain model behavior to non-ML stakeholders.
Passion for machine learning and its real-world applications, particularly state-of-the-art neural networks.

Nice To Haves

Experience with vision-language models such as CLIP or LLaVA-style models, or open-vocabulary detection.
Experience with multi-object tracking and video analytics, including re-identification, action recognition, or temporal models.
Experience with few-shot learning, active learning, or synthetic data generation for low-data domains.
Familiarity with MLOps tooling such as MLflow or Weights & Biases, Kubernetes, Triton Inference Server, or CI/CD for models.
Domain experience in healthcare, retail and supply chain, manufacturing inspection, workplace safety, or video surveillance.

Responsibilities

Design, train, and fine-tune models for object detection, classification, segmentation, tracking, and anomaly detection using customer video and images.
Adapt pretrained and foundation models to new customer use cases, including use cases with limited labeled data.
Own data pipelines, including data curation, labeling guidelines, augmentation, synthetic data, and dataset versioning.
Optimize models for production and edge hardware using techniques such as quantization, pruning, distillation, ONNX export, and TensorRT to meet latency and accuracy targets.
Build evaluation suites with clear metrics and identify regressions before release.
Deploy and monitor production models, diagnose drift and failure cases, and support retraining.
Work with product teams, solutions engineers, and customers to translate requirements into measurable model specifications.
Write clean, tested, reviewed code and contribute to shared machine learning tooling.
Collaborate on collecting, processing, and maintaining high-quality datasets for model training and testing.
Implement computer vision algorithms and integrate machine learning models into broader products and systems.
Evaluate model performance using qualitative and quantitative methods and iterate to improve accuracy and efficiency.
Stay informed about developments in machine learning and computer vision and apply relevant approaches to technical challenges.
Engage internal and external stakeholders to understand their needs and develop technical solutions.
Troubleshoot problems arising during computer vision system development and deployment.
Travel on site when needed to assess environments, test systems in real-world conditions, and verify operation.