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 owns the complete toll collection and highway safety system for toll authorities and state DOTs across the US Interstate Highway System. The company processes 11 million images daily using computer vision (detection, segmentation, classification, embeddings, re-identification) to match vehicles to tolls. With fewer than 30 employees, Blissway operates a lean, high-impact team in a roughly $15B global tolling industry expected to double within the next decade.
In this role, you will own the machine learning operations infrastructure that takes models from development to production and keeps them accurate and reliable at scale. You'll be responsible for the complete pipeline: training infrastructure, model serving, versioning, deployment, rollback, and monitoring across both cloud and edge devices. Your work directly impacts the roadside hardware deployed on major highways nationwide.
Key responsibilities include:
- Building and maintaining the infrastructure behind every model shipped to production
- Optimizing vision pipelines and inference paths for 11 million daily images
- Designing systems for edge deployment on resource-constrained roadside hardware
- Implementing monitoring and observability to catch model degradation before it affects toll accuracy
- Owning model updates and retraining workflows that keep the system correct in the field
- Collaborating across a small, adaptable team where you'll help other teams when needed
The company runs its own servers rather than relying on fully managed cloud services, giving you direct control over infrastructure. You'll work with real hardware in the real world, not just test environments. The tech stack is primarily Python and TypeScript, though language proficiency is less important than the ability to learn quickly and write production-quality code.
Blissway emphasizes data privacy and legal compliance: the company works only for toll authorities and state DOTs on legislatively authorized toll roads. Laws governing the work prohibit keeping data unless used to charge a toll—there is no search-by-plate interface, cross-agency lookup, or law enforcement access.
The team expects high ownership and commitment: recent surveys show an average of 55 hours per week, with occasional 70+ hour bursts for major releases. You should be detail-obsessed, adaptable, and comfortable with the grind of shipping systems that work reliably in the unpredictable physical world.
REQUIREMENTS:
- 2 to 6 years of software engineering experience building and running production systems (infrastructure, backend services, or ML platforms)
- Production fundamentals: ability to write production-quality code and build reliable systems including pipelines, deployment, observability, and operational tooling
- ML curiosity: understanding of how models are trained and evaluated well enough to debug them in production; prior model deployment experience is helpful but not required
- Bonus qualifications: experience taking ML models from notebook to production (deployment, monitoring, retraining), deploying to edge or resource-constrained hardware, or running self-managed servers instead of fully managed cloud services
