Principal Product Manager, Model Customization & Open-Weight Models

Equinix•Toronto, ON

3d•$190,000 - $286,000

About The Position

Equinix is hiring a Principal Product Manager to own the strategy and lifecycle of model selection and customization products. The role will define how Equinix selects models for different applications, maintain a recommended-model catalogue covering open-weight and proprietary models, and benchmark new releases against Equinix workloads. It will establish practical approaches to model adaptation, choosing among prompt and context engineering, retrieval, fine-tuning, preference optimization, reinforcement learning, distillation and quantization according to demonstrated value. The position owns the roadmap for a shared customization platform spanning data preparation, training, evaluation, model registry, deployment and lifecycle management. It partners with infrastructure and engineering teams on GPU capacity, training and serving systems, inference optimization and cost controls, and ensures tuned models are reproducible and supportable. The role also works with business and domain teams to take use cases from scoping and experimentation through production launch, and converts successful approaches into reusable patterns and services. The Principal Product Manager will define model evaluation and launch standards, measure production results, and establish governance for licensing, provenance, data rights, documentation, red-teaming and deployment approval. As a cross-organizational product leader, the role will align engineering, data science, infrastructure, security and business leaders through strategy, specifications and executive communications. Success includes publishing a benchmark-backed model selection framework and catalogue, launching a shared customization platform, deploying tuned open-weight models with at least two vertical teams with measured gains, and putting an efficient governance process in place.

Requirements

At least 7 years of experience across product management, engineering or applied machine learning, including at least 3 years as a product manager on AI/ML, data or infrastructure platforms.
Hands-on experience customizing and shipping open-weight models in production, such as through fine-tuning, parameter-efficient tuning, preference optimization, distillation or quantization.
Practical knowledge of prompt and context engineering and retrieval-augmented generation, with sound judgment about when to use these methods instead of training.
Experience selecting models for real applications while weighing quality, cost, latency, hardware and licensing.
Strong evaluation skills, including building task-specific evaluations, avoiding train/test leakage and distinguishing genuine improvement from noise.
Working knowledge of model training and serving stacks, such as PyTorch, Hugging Face and inference servers such as vLLM or SGLang, as well as GPU cost and capacity trade-offs.
Ability to read and write Python, run experiments and review training and evaluation results directly.
Ability to lead across engineering, science, infrastructure and business teams without formal authority and to explain trade-offs clearly to senior leaders.

Nice To Haves

Experience with reinforcement learning from human or AI feedback, or reward modeling.
Experience with synthetic data generation and data curation at scale.
Experience with tool-use or agentic fine-tuning, or function-calling models.
Experience with open-weight model licensing, model governance or responsible AI programs.
Experience running models in private, hybrid or on-premises environments, or working with data residency requirements.
Experience in regulated industries such as financial services.
A degree in computer science, engineering, machine learning or another quantitative field, or equivalent experience.
Publications or open-source contributions.

Responsibilities

Define model-selection criteria for Equinix applications, including task quality, cost, latency, context length, tool-use ability, hardware footprint, license terms and data-handling requirements.
Maintain a recommended-model catalogue covering open-weight and proprietary models, with guidance on appropriate use cases.
Track model releases, benchmark them against Equinix workloads and determine when upgrades justify switching costs.
Set the playbook for model adaptation, including prompting, context engineering, retrieval, supervised fine-tuning, parameter-efficient methods, preference optimization, reinforcement learning from feedback, distillation and quantization.
Guide teams toward the lightest model customization technique that meets requirements, using prompting and retrieval before training where appropriate.
Define data strategies for tuning, including curation, labeling, synthetic data generation, deduplication and prevention of training/evaluation leakage.
Own the roadmap for a shared model customization platform covering data preparation, training, evaluation, registry, deployment and lifecycle management.
Partner with infrastructure and engineering teams on GPU capacity, training and serving stacks, inference optimization and cost controls.
Ensure tuned models are reproducible and supportable through versioned data, configurations and checkpoints, clear ownership and rollback processes.
Work with business and domain teams to scope use cases, select models, design tuning experiments and launch solutions to production.
Turn successful early deployments into reusable patterns, templates and services.
Conduct model launch reviews covering quality, safety, cost, performance and operational readiness.
Define evaluations that models must pass before launch, including task-specific quality, regression against the base model, safety, robustness, cost and latency targets.
Partner with the Agent Intelligence team to measure tuned models in production and feed results back into improved data and models.
Report customization impact through measures such as quality gains, cost savings, latency reductions and risk avoided.
Set standards for license review, model provenance, data rights, documentation, red-teaming and deployment approval.
Ensure confidential and customer data used in tuning is handled in accordance with Equinix security, privacy and regulatory requirements.
Serve as Equinix’s product leader for model selection and customization, influencing teams without direct authority.
Write strategy, specifications and executive briefs to align engineering, data science, infrastructure, security and business leaders.
In the first year, publish a benchmark-backed model selection framework and catalogue, launch a shared customization platform, ship tuned open-weight models with at least two vertical teams with measured baseline gains, and establish an efficient governance process.