SlipstreamJobs tracks this role from the company's public career site. Apply directly on the employer's site.
Stream (GetStream.io) is seeking a Staff AI Engineer to own end-to-end model development on its AI team. You will build, fine-tune, evaluate, and ship production models that power real-time communication and social experiences across Stream's products, which serve over a billion end users globally. Stream is known for its SDKs and APIs enabling activity feeds, chat, and voice/video capabilities, as well as Vision Agents, an open-source Python framework for building low-latency AI agents.
In this role, you will own the complete lifecycle of in-house AI models: dataset design, supervised fine-tuning, post-training experiments, evaluation harness development, data pipeline construction, and production deployment. You'll establish benchmarks to determine model readiness, maintain high-quality data pipelines for training, optimize models for latency and cost at scale, and set technical direction for undefined problem spaces. You'll work across the engineering organization, interfacing with Go-based API teams and infrastructure to integrate models into products. You're expected to contribute to the open-source ecosystem and raise engineering standards through code review and mentorship.
Required qualifications include 5+ years of production Python engineering with shipped, maintained code; hands-on machine learning experience with supervised fine-tuning and post-training; familiarity with modern fine-tuning and serving tools (Unsloth, Fireworks, Baseten); cloud experience with GCP or AWS including infrastructure-as-code; proven track record running ML-based products in production (deployment, monitoring, retraining, iteration); experience designing and operating data pipelines; demonstrated ownership of ambiguous problems; and strong communication skills in distributed, fast-moving teams.
Preferred qualifications include visible open-source contributions, Go experience, deep Python concurrency knowledge, real-time/low-latency inference experience, and early-stage startup experience.
The role is based in Amsterdam with a hybrid policy requiring 3 days per week in office for Netherlands-based applicants, with exemptions available. Remote work across Europe is available for eligible candidates. Occasional travel for in-person collaboration is encouraged. Stream offers 28 days PTO, company equity, pension, L&D budget, relocation support, and Amsterdam office perks.
