About DaoAI
DaoAI Robotics builds AI-native Automated Optical Inspection (AOI) systems for SMT/EMS electronics manufacturing. Our SMT AOI system is already running on customer production lines, inspecting component placement and solder quality inline on PCBA lines. We're in a phase of fast iteration, and we're growing the team. We fully embrace AI-assisted development: the team uses code agents such as Claude Code and Codex every day.
The Role
You'll join the vision algorithm team and work on real production data and real inspection problems, with a senior engineer as your mentor. You start with well-scoped pieces of the inspection pipeline - data, training, evaluation, failure analysis - and take on more ownership as you grow.
This is a full-time role for new graduates and engineers with 0–2 years of experience. We expect you to grow into owning a full inspection capability within your first year.
What You'll Do
- Implement and improve PCB/PCBA inspection algorithms in one or more areas: component detection and localization, segmentation, defect classification, character recognition, anomaly detection.
- Build and maintain datasets: curate and clean production images, apply labeling rules, and review label quality.
- Train and evaluate models, run controlled experiments, and report results clearly with numbers.
- Analyze false calls and escapes from customer lines, find the root cause, and turn each one into a fix or a regression test.
- Write classical image-processing code (alignment, template matching, measurement) alongside deep-learning models, and build small tools for visualization and evaluation.
- Work with the software team on model export, integration, validation and field-issue debugging.
What You Bring
- Bachelor's degree or higher in Computer Science, Electrical/Computer Engineering, Mathematics or a related field. New graduates and candidates with 0–2 years of experience are welcome.
- Solid deep-learning vision fundamentals and at least one project where you trained and evaluated a model yourself in PyTorch (detection, segmentation, classification or anomaly detection) - coursework, thesis, internship or industry all count.
- Working knowledge of classical computer vision with OpenCV: filtering and morphology, template matching, geometric transforms, basic fitting.
- Evaluation discipline: you understand precision and recall, know how to build a trustworthy test set, and can explain why a model got a sample wrong.
- Good Python and Git habits; hands-on with AI coding tools (Claude Code, Codex, Cursor, etc.).
- Curiosity and ownership: you dig into failure cases until you understand them, and you check your numbers before you report them.
Nice to Have
- Exposure to industrial inspection, PCB/PCBA, AOI/SPI or SMT processes and common defects (insufficient/excess solder, bridging, tombstoning, misalignment, missing parts, reversed polarity).
- Deployment experience: ONNX, TensorRT, quantization, or C++ inference.
- Experience with OCR, few-shot/unsupervised anomaly detection, or vision foundation models.
- Publications, open-source contributions or competition results.
Why Join
- Your models run on real factory lines, and you see the impact directly.
- Direct mentorship from a small, senior team with an AI-first engineering workflow.
- Exposure to the full stack: deep learning, classical vision and production systems.
Salary: CA$60,000 per year. Location: Vancouver, BC (Hybrid). Full-time.