SlipstreamJobs tracks this role from the company's public career site. Apply directly on the employer's site.
Blissway is a deep-tech startup that simplifies toll collection and improves highway safety through AI/ML, hardware, SaaS, and IoT. The company processes 11 million images daily across the Interstate Highway System and operates with a lean team of fewer than 30 people.
As a Machine Learning Engineer, you will own the complete ML pipeline from raw sensor data to production inference. This is not a research role—you ship models to real roads. Your responsibilities include:
- End-to-end ownership: dataset curation, model training, deployment, monitoring, and iteration
- Working with real hardware deployed in the field, enabling rapid testing and validation on actual roads
- Processing computer vision at scale: detection, segmentation, classification, embeddings, and re-identification across vehicle detection, license plates, wheels, lane markings, and multi-sensor data
- Selecting and implementing the right tools, from classical computer vision algorithms to custom state-of-the-art models
- Balancing cloud and edge inference: optimizing models for both unlimited cloud compute and resource-constrained roadside hardware (speed, size, power efficiency)
- Staying current with ML research and holding work to frontier standards
The company emphasizes a lean, adaptable team culture. You will have direct access to technical leadership, real autonomy from day one, and work that ships to production in weeks rather than quarters. The role involves occasional 70+ hour weeks during major releases (average 55 hours/week).
Requirements:
- 2 to 6 years of software engineering with focus on machine learning and/or computer vision
- True end-to-end experience: models taken from raw data to production with ownership of post-deployment performance
- Strong software engineering fundamentals plus hands-on ML expertise; comfortable writing production-quality code and training/debugging models
- Full lifecycle ownership: trained, deployed, monitored, and iterated models in production
- Real computer vision depth with judgment to apply classical techniques versus heavy models appropriately
- Proficiency in Python and/or TypeScript preferred, though fast learning in any language is acceptable
