Staff Backend Engineer – AI at Stream

on-site · full_time · Visa sponsorship

Apply for this role at Stream

Responsibilities:
- Own end-to-end model development: dataset design, supervised fine-tuning, post-training experiments and evaluation harnesses.
- Build and maintain high-quality data pipelines for training, labelling and reproducibility.
- Deploy models to production and optimise serving for latency, cost and reliability at scale.
- Define technical direction for ambiguous problem spaces and make pragmatic trade-offs on what to build and ship.
- Collaborate with API, Go, and infrastructure teams to integrate models into product systems and global edge serving.
- Raise engineering standards via code review, mentorship and sharing work with the community/open-source when relevant.

Requirements:
- 5+ years of production-level Python engineering with shipped, maintained code (not just prototypes).
- Hands-on ML experience, specifically supervised fine-tuning and post-training workflows.
- Familiarity with modern fine-tuning and serving toolchains (e.g., Baseten, Fireworks or equivalents).
- Cloud experience (GCP or AWS) and infrastructure-as-code (Terraform) for deployment and infra management.
- Experience running models in production: monitoring, retraining, latency tuning and cost/reliability trade-offs.
- Strong systems and distributed-systems experience for low-latency, high-volume inference pipelines.