Jobs at Boson AI

Eight-month applied research internship for University of Toronto MScAC students. Interns will work with Boson AI mentors on speech, audio, video generation, computer vision, efficient training, real-time inference, conversational agents, reinforcement learning, or full-duplex multimodal conversation. Responsibilities include building data pipelines, designing evaluations, developing and optimizing models, and conducting rigorous experiments on large-scale GPU infrastructure.

Own Boson AI’s developer relations program, helping developers adopt its real-time speech and multimodal AI products. Build tutorials, notebooks, sample applications, integrations, demos, and launch content; support open-source communities; present publicly; triage developer issues; and translate feedback into product improvements. The role requires hands-on Python and AI API development, technical communication, ecosystem management, and up to 20% travel.

Build and operate the core platform for Boson AI's model APIs and agentic products: API serving, state management, distributed services, data pipelines, SDKs, retrieval and caching, observability, and integrations with ML and product teams. Grow ownership under senior mentorship while shipping production features.

Lead development of AI-powered enterprise platforms, focusing on full-stack web development, UI/UX design, and data visualization.

Design, operate, and automate reliable infrastructure for large-scale AI training and inference. Own networking, compute, storage, GPU/server configuration, and operational tooling. Build monitoring, alerting, runbooks, incident response, capacity planning, provisioning automation, and collaborate with ML and platform teams to improve performance and reliability.

Develop and deploy AI solutions for enterprise clients, focusing on large language models and agentic systems integration while ensuring quality, scalability, and security.

Own and evolve the core platform for model APIs and agentic products: build high-throughput, low-latency distributed services, data pipelines, SDKs, context retrieval, observability, and collaborate with ML/product teams to meet strict latency and reliability SLIs.