Paper is building the first B2B, in-classroom Voice + Video AI Tutor designed for real-world learning environments. They are seeking a Senior AI Engineer to architect and operationalize multimodal AI systems at production scale, owning end-to-end AI initiatives that power features used by millions of learners.
Responsibilities:
- Design and deploy real-time voice and vision AI systems for classroom environments
- Evaluate tradeoffs between fine-tuning, RAG pipelines, prompt-based approaches, and tool-using agents
- Build scalable inference systems with strict latency and cost constraints
- Fine-tune and evaluate LLMs, ASR models, and vision/multimodal architectures
- Design robust evaluation frameworks (offline + online)
- Implement safety guardrails and bias mitigation strategies
- Own the full ML lifecycle: data ingestion, training, validation, deployment, monitoring, and iteration
- Optimize infrastructure for GPU utilization, latency, and cost efficiency
- Partner with MLOps to build reliable, versioned AI pipelines
- Work closely with Product and Engineering to translate classroom needs into production-ready AI systems
- Clearly communicate tradeoffs between model complexity, performance, and cost
- Contribute to long-term AI roadmap and architecture decisions
Requirements:
- 5+ years building and deploying production AI/ML systems
- Hands-on experience operationalizing LLM-based systems (RAG, agents, prompt workflows, fine-tuning)
- Experience with multimodal or vision systems
- Demonstrated ownership of AI systems in production (monitoring, evaluation, iteration)
- Strong Python proficiency and experience with PyTorch or TensorFlow
- Experience working with distributed/cloud infrastructure for ML workloads
- Experience building real-time or low-latency ML systems
- Experience working with educational data or regulated environments (FERPA/COPPA)
- Experience designing evaluation frameworks for generative systems
- Background in optimizing ML systems for cost at scale