Who We Are
Morningside AI builds and implements AI for companies that can't afford to get it wrong. Fortune 500s, NBA teams, major media groups, they come to us because we actually ship, not just strategize.
We started as a development studio. Now we're a full transformation partner: development, consulting, education. Whatever the problem requires. We've spent the last two and a half years figuring out what actually works when you put AI into production - not theory, not demos, real systems that run real businesses.
We're headquartered in New Zealand with team members working remotely across the world.
Morningside AI Is Not a Typical Development Agency. We Believe...
- Perfection is a trap. We build, deploy, learn, and improve. If you need six months of planning before you can start, this isn't the place.
- No layers of approval, no death by meeting, no waiting for permission. You own your work from day one, which means real accountability too.
- No politics, no sugar-coating, no guessing games. We give direct feedback and we expect the same. Problems get named so they can get fixed.
- We use AI to build AI. If you're still writing every line by hand, you're already behind. We use every tool available to move faster than teams twice our size.
- We don't care when you're online. We care about what you ship.
The Role
We're looking for a part-time Computer Vision Engineer (16-24 hours per week) to work alongside our engineering team on a growing portfolio of CV projects. This is an initial engagement of a couple of months with strong potential to continue for the right person - we have consistent and expanding CV work and want someone who can grow with us.
You won't be handed a fully scoped ticket. You'll be working on hard problems with real-world data, and you'll need to make good judgments about where to spend time and when to pivot.
What You'll Work On
- Object detection and counting pipelines on complex, domain-specific image inputs
- Evaluating and combining classical CV, few-shot detection, and fine-tuned model approaches depending on what the data and constraints call for
- Tiled inference pipelines to handle large, high-resolution images
- Annotation workflows and custom model training on domain-specific datasets
- Accuracy evaluation, failure mode analysis, and documentation
- Supporting LLM/VLM integration for classification, edge case handling, and verification layers where relevant
Must Haves
- Hands-on production experience with computer vision — object detection, classification, counting, or segmentation using OpenCV, PyTorch, or similar
- Experience with detection architectures like YOLO, Faster R-CNN, or Mask R-CNN, and the judgment to choose between them based on the problem
- Comfortable working with domain-specific or limited training data — you know how to get results without 10k labeled examples
- Python fluency across the full ML lifecycle: data prep, training, evaluation, inference
- You work independently, flag blockers early, and don't need constant direction
Nice To Haves
- Experience with few-shot or zero-shot detection approaches (T-Rex2, GeCo, or similar)
- Familiarity with tiled inference strategies for large image inputs
- Experience with synthetic data generation for model training
- LLM/VLM integration experience (Claude, GPT-4o, or similar) as a supporting layer alongside CV pipelines
- Prior work on technical drawings, scanned documents, or other non-photographic image inputs
- Experience with annotation tooling and managing labeling workflows
- Generative AI or RAG pipeline experience
- MLOps or model serving experience — containerizing models, deploying on AWS/GCP, managing inference at scale
- Consulting or agency background where you've had to explain technical tradeoffs to non-technical stakeholders
The Details
- Part-time: 16-24 hours per week, flexible scheduling
- Remote, async-first
- Initial engagement of a couple of months, with strong likelihood of ongoing work
- Competitive hourly rate