SlipstreamJobs tracks this role from the company's public career site. Apply directly on the employer's site.
Twelve Labs builds the intelligence layer for video understanding using multimodal AI models that comprehend video across sight, sound, and motion. The company powers production-scale AI workloads across media, entertainment, sports, security, and government, backed by $210M+ from NEA, Radical Ventures, Amazon, NVIDIA, Snowflake, and others.
The Cognition Models team owns the models that turn video into structured understanding and reasoning, including Pegasus (video-language model) and Jockey Core (reasoning LLM). Pegasus is a video-language model that reasons over visuals, speech, audio, and on-screen text to produce time-based, structured metadata—enabling customers to define segment types and metadata fields they care about, with Pegasus returning relevant start/end times and structured outputs for search, archive, editing, compliance, and content-management workflows.
In this role, you will build, improve, and operate production ML systems for Pegasus with a focus on reliability, performance, and maintainability. You'll work across core parts of the ML stack including deployment, inference, evaluation, monitoring, and supporting infrastructure. Key responsibilities include developing systems for serving Video Language Models (VLMs) and handling multimodal data and metadata at production quality, making strong technical decisions within your area, and driving execution with high ownership. You'll explore and adopt AI-assisted development tools (Claude, Gemini, GPT) to improve productivity across coding, experimentation, debugging, and documentation.
You should have strong software engineering and machine learning fundamentals, experience building and shipping ML systems in production, and familiarity with multimodal data (computer vision, NLP, LLMs, or VLMs). Experience with distributed ML or data workflows in Kubernetes-based environments and strong engineering judgment around performance, reliability, and maintainability in production are highly valued.
