Build and maintain production ML infrastructure for a toll collection system processing 11 million images daily. Own model deployment, monitoring, and edge-to-cloud pipelines that keep vision models accurate in real-world conditions. Requires 2-6 years of software engineering experience with Python or TypeScript and a strong interest in ML systems.
Key Highlights
Process 11 million images per day using detection, segmentation, classification, embeddings, and re-identification for accurate toll transactions
Own the full ML lifecycle: training/serving infrastructure, model versioning, deployment, rollback, monitoring, and updates
Deploy models to both cloud and edge hardware in real-world roadside environments
Lean startup team of fewer than 30 with an average 55-hour work week and occasional 70+ hour bursts
Based in Denver, Colorado with in-person collaboration required
Key Responsibilities
Own the infrastructure behind every model shipped including training and serving infrastructure, model versioning, deployment, and rollback
Build and maintain vision pipelines processing 11 million images per day across vehicles, license plates, wheels, and lane markings
Deploy models to edge and cloud environments, building paths that get models onto roadside hardware and keep them updated
Build monitoring systems that catch model degradation before it impacts toll transactions
Own model updates that fix accuracy issues in production
Discover our full range of relocation jobs with comprehensive support packages to help you relocate and settle in your new location.
Benefits & Perks
Relocation bonus
Personalized health coverage (ICHRA) with monthly allowance
401(k) matching up to 4%
Company-sponsored life and disability insurance
Early stage equity with annual tender process
4 weeks of untracked PTO
12 weeks paid parental leave
12 weeks fully paid sabbatical every 5 years
Daily group lunches and stocked kitchen
Annual 4-day team getaway (BlissTrip)
Tuition reimbursement for courses, programs, and conferences
Nice to Have
Experience taking ML models past the notebook including deployment, monitoring, and retraining
Experience deploying to edge or resource-constrained hardware
Experience running your own servers instead of fully managed cloud services
