Role overview
Gradient is a small business in Renton, WA. Our goal is to provide real world applications utilizing AI and robotics. We apply AI and robotics to solve classification, identification, and quality assessment challenges across physical product domains. Come join us to work on challenging problems, serve real customers, and gain experience with a small startup.
We are looking for an AI Engineer to join our team and help develop next-generation Computer Vision and Language Model systems for real-world applications. The ideal candidate will have hands-on experience with image segmentation, vector similarity search, and LLM-based feature extraction. You’ll work on training and optimizing deep learning models that detect fine-grained visual or textual patterns and contribute to building scalable AI infrastructure.
This is an excellent opportunity for an individual passionate about practical AI and eager to collaborate in a fast-paced start-up environment alongside experienced engineers and researchers.
What you'll work on
- Develop and train segmentation and defect detection models for identifying visual features such as surface wear, shape anomalies, or alignment inconsistencies.
- Build and optimize image embedding and vector similarity search pipelines (e.g., using FAISS, Milvus, or Pinecone) for matching against large reference datasets.
- Experiment with pattern recognition and feature consistency methods to identify relationships between visually similar items or product variants.
- Apply LLM-based feature extraction to interpret structured and unstructured text—automatically deriving attributes like names, identifiers, years, and categorical fields from images or metadata.
- Design and build end-to-end data extraction pipelines that parse unstructured sources into clean, validated, production-ready datasets. This includes schema design, multi-pass parsing with edge case handling, cross-column validation, and iterative QA cycles to ensure data integrity at scale.
- Evaluate and fine-tune model performance across multiple architectures and modalities (vision, text, or multimodal).
- Collaborate closely with software and data engineers to integrate AI models into production pipelines and APIs.
- Research and prototype state-of-the-art AI techniques in segmentation, embedding learning, and multimodal understanding.
- Maintain thorough documentation of experiments, models, and deployment results.
What we're looking for
- Currently pursuing or holding a degree in Computer Science, Electrical Engineering, Artificial Intelligence, or related fields.
- Strong programming skills in Python with experience using PyTorch, TensorFlow, or JAX.
- Familiarity with Computer Vision tasks such as segmentation, detection, and classification using frameworks like OpenCV, MMDetection, or Detectron2.
- Understanding of embedding spaces, vector databases, and similarity search techniques (FAISS, Annoy, Milvus, Pinecone, etc.).
- Experience fine-tuning or prompting Large Language Models (LLMs) for structured information extraction or classification tasks.
- Exposure to multimodal models (e.g., CLIP, BLIP, Florence, or Segment Anything Model).
- Proficiency in data wrangling and labeling pipelines, including dataset curation and annotation tools.
- Docker and container orchestration
- Strong analytical and problem-solving abilities, with attention to data quality and reproducibility.
- Excellent collaboration and communication skills across multidisciplinary teams.
- A self-starter mindset with curiosity for exploring emerging AI research and applying it in production environments.
Pay: $85,000.00 - $200,000.00 per year
Benefits:
- Health insurance
- Paid time off
Ability to Relocate:
- Renton, WA 98057: Relocate before starting work (Preferred)
Work Location: In person