Project Manager - Higgs Audio

Boson AI•Santa Clara HQ, CA

10h•Hybrid

About The Position

Boson AI is hiring a Project Manager to own Higgs Audio releases and roadmap, guiding work from research through launch. Higgs Audio is a family of foundation audio and avatar models that generate and recognize speech, clone voices, detect sentiment, and provide voices with faces. The role coordinates researchers, engineers, customers, partners, and product teams to ship releases with clear quality standards and predictable schedules. This is a hands-on senior individual-contributor role. The Project Manager does not need to train models but should be technically fluent enough to reproduce API bugs, interpret evaluation reports, listen critically to generated speech, and distinguish real regressions from noise. The role is based at either Boson AI’s Santa Clara headquarters or Toronto office, on a hybrid schedule requiring at least three office days per week. Occasional travel between the offices and to customers is expected.

Requirements

At least 5 years of project, program, or product management experience on software products, including at least 2 years working on technical products such as APIs, developer platforms, or ML models.
Track record of shipping on schedule, including running release processes and launch checklists and making go/no-go decisions.
Technical fluency with Python, REST and streaming APIs, GitHub, and issue trackers; ability to reproduce bugs, read logs, and analyze evaluation results using SQL, spreadsheets, or notebooks.
Experience translating customer conversations and usage data into prioritized requirements, with the judgment to decline requests when appropriate.
Clear, concise writing for specifications, release notes, and status updates.
Ability to use agentic AI efficiently in work.
Ability to work from the Santa Clara headquarters or Toronto office on a hybrid schedule.

Nice To Haves

Experience with speech or audio products such as TTS, ASR, voice cloning, voice agents, or audio for media production.
Experience releasing ML models, including evaluation suites, human preference studies, model cards, and open-weight releases on Hugging Face.
Experience managing issues, pull requests, and releases for an open-source project on GitHub.
Experience with enterprise customers or cloud partners such as Microsoft Foundry, AWS Bedrock, or Google Vertex AI.
A good ear for identifying and describing problems with generated voices.
An engineering background or a degree in computer science, electrical engineering, or a related field.
Fluency in a language besides English.

Responsibilities

Own Higgs Audio releases from planning through launch, including schedules, go/no-go reviews, documentation, release notes, model cards, and changelog entries.
Coordinate releases across the API and Boson Workspace, Hugging Face open weights, GitHub repositories, and hosting partners such as Microsoft Foundry and Deep Infra.
Work with research and engineering to define and enforce release quality standards, including multilingual error rates, voice-cloning similarity, expressiveness benchmarks, listening tests, latency, and cost.
Triage bug reports and feature requests from GitHub, customer support, partners, and internal teams; assign severity, priority, and owners; and close the loop with reporters.
Break goals into milestones, track dependencies across research, engineering, and go-to-market, raise risks early, and make tradeoffs explicit.
Engage developers, enterprise customers, partners, and creators; run beta programs; review usage data and support tickets; and translate findings into requirements and priorities.
Ensure model changes integrate smoothly into products built on them, such as Higgs VoiceOver.
Own versioning and deprecation, including version names, migration guides, and advance deprecation notices for breaking changes.
Track release cadence, customer-impacting regressions, critical-issue resolution time, adoption, and customer satisfaction; provide regular status updates to leadership.
Establish a predictable published release schedule and documented quality bar, ensure open issues have owners and priorities, and drive timely resolution of critical bugs.
Maintain a public changelog for model and API changes, provide advance notice of breaking changes, and use regular customer reporting to inform the roadmap.