About Wingfin
Wingfin develops autonomous driving systems. Our mission is rooted in accessibility: human-like autonomy for safe, cost-efficient, real-world scalability. By leveraging cutting-edge deep learning and end-to-end neural networks, Wingfin is redefining the landscape of self-driving technology. We are building a practical, cost-efficient platform that brings safe autonomy to every vehicle on the road. Headquartered in Singapore, the company boasts an international team of multi-talented experts. We are looking for talented and motivated individuals to join our team either in Singapore or Malaysia.
The Role
We are looking for an entry-level Machine Learning Engineer with a strong quantitative background who enjoys building things, running experiments, working with real data, and figuring out why things do or do not work. A Bachelor's degree is sufficient, but we are looking for candidates with a strong academic foundation.
We value curiosity, practical ability and hands-on experimentation. Candidates should have a strong academic foundation in Physics, Mathematics or Engineering.
Our ML work spans perception, vehicle control and planning, using large-scale real-world and simulated driving datasets. Your initial focus will be on preparing, processing and validating data for ML training, with opportunities to become increasingly involved in model development and experimentation.
What You Will Work On
- Prepare and process large autonomous-driving datasets for ML training.
- Build pipelines to extract, clean, filter, label and organise camera, vehicle-state and other sensor data.
- Inspect datasets for errors, inconsistencies, unusual cases and data-quality problems.
- Develop tools for automatically identifying useful driving scenarios and training examples.
- Validate datasets before they enter model-training pipelines.
- Run ML experiments and analyse their results.
- Work with engineers developing models for perception, planning and control.
- Help improve training and evaluation infrastructure as our datasets and models grow.
- Investigate failures and unusual model behaviour by tracing them through models, code and underlying data.
Core ML areas
Perception: understanding roads, vehicles, pedestrians, traffic infrastructure and the driving environment.
Planning: predicting and selecting appropriate trajectories and driving behaviour.
Control: translating desired behaviour into steering, acceleration and braking.
What We Are Looking For
- Bachelor's degree or higher in Physics, Mathematics or Engineering, with a strong academic record.
- Some practical experience with machine learning. This does not need to come from a commercial job - university work, research, personal projects, competitions or substantial independent experimentation are all valid.
- Use of modern AI tools is a requirement. We expect active use of AI-assisted tools for coding, debugging, research, data analysis and day-to-day engineering productivity.
- Comfortable programming, preferably in Python.
- Good understanding of basic ML concepts such as training/validation datasets, loss functions, overfitting and model evaluation.
- Strong numerical and analytical ability.
- A willingness to work directly with data rather than treating datasets as a black box.
- Excellent attention to detail. Small errors in timestamps, labels, coordinate systems or dataset construction can invalidate an otherwise good experiment.
- Able to investigate problems systematically and independently.
- Strong preference for hands-on experimentation: we want people who naturally try things, measure what happens and iterate.
Useful, But Not Required
Experience with any of the following would be helpful, but we do not expect an entry-level candidate to know all of them:
- PyTorch or another modern deep-learning framework
- Computer vision
- Object detection, segmentation or tracking
- Transformers or modern neural architectures
- Robotics or autonomous systems
- Vehicle dynamics or control systems
- ROS / ROS2
- Large image or video datasets
- Linux and command-line development
- Git
- CUDA / GPU-based training
- Simulation environments such as CARLA
The Kind Of Person We Want
You might be a good fit if you have ever:
- Trained a model simply because you wanted to see whether an idea would work.
- Written scripts to collect or clean your own dataset.
- Built a computer-vision, robotics or ML project outside formal coursework.
- Found yourself digging through data because a model result "did not look right."
- Compared several approaches experimentally rather than relying only on what a textbook said should happen.
What We Offer
- The opportunity to work directly on real autonomous-driving ML systems.
- Exposure to perception, planning and control rather than a narrowly defined ML task.
- Large real-world and simulated driving datasets.
- Significant scope to learn and take on model-development responsibilities as you grow.
- A technically demanding environment where good experiments and good engineering matter more than titles.
- Competitive salary and benefits.
This role is suitable for a strong new graduate or someone in the first few years of their career who wants to build a serious foundation in applied machine learning and autonomous systems.