Member of Product Staff, Foundation Models & Post-Training

Abridge

Where
Hybrid in SF Office
Type
Full time

Posted 6 days ago· Checked live at Abridge 22 minutes ago

Live.Open on the employer's own site.

See how you score for this job

One sentence and your resume. We score it against what Abridge asks for, line by line.

See my score
Is this job still open?
Yes. We checked it live on Abridge's own hiring page 22 minutes ago.
What is the job?
Founded in 2018, Abridge is the first AI-native clinician intelligence platform organized around the patient, built for clinicians, and designed to help health systems coordinate the clinical, financial, and evidence-based decisions for every moment of care. Powered by purpose-built AI for healthcare, Abridge connects patients, providers, payers, and life sciences companies.
What does it pay?
The posting doesn't say.
Our verdict
Live. Open on the employer's own site.

What the job is

Founded in 2018, Abridge is the first AI-native clinician intelligence platform organized around the patient, built for clinicians, and designed to help health systems coordinate the clinical, financial, and evidence-based decisions for every moment of care. Powered by purpose-built AI for healthcare, Abridge connects patients, providers, payers, and life sciences companies.

What you'll do

  • Set the product strategy for Abridge's model family. Translate company and product priorities into a coherent portfolio of models, with explicit choices across capability, quality, latency, cost, safety, and controllability.
  • Own the path from research to product impact. Define the hypotheses, milestones, decision gates, and success metrics that move models from experiments into shadow mode and production.
  • Turn proprietary data into a product advantage. Shape how Abridge uses de-identified conversations, final notes, clinician edits, EHR context, and care actions as training and feedback signals, in partnership with Data, Privacy, Security, and Clinical teams.
  • Make model investment decisions legible. Build clear frameworks for when to use a frontier model, an open model, a prompted workflow, or an Abridge-trained model; quantify expected quality, serving-cost, latency, control, and strategic benefits.
  • Partner with Evals on promotion criteria. Define the product-relevant capabilities and failure modes that matter for each model use case, while the Evals team owns shared measurement infrastructure and neutral evaluation standards.
  • Create a tight learning loop with product teams. Convert production failures, clinician feedback, edit behavior, and emerging product needs into training priorities; make model improvements visible in user and business outcomes.
  • Drive cross-functional execution. Align ML Science, ML Engineering, Product Engineering, Clinical Science, Data, Evals, and product pods around priorities, interfaces, ownership, and delivery - often without formal authority.
  • Communicate the strategy. Make complex research choices understandable to executives and product teams, clearly separating demonstrated results, working hypotheses, and long-term bets.

What you bring

  • 7+ years of product-management or closely related experience, including substantial ownership of ML-powered products, model platforms, or AI infrastructure.
  • A track record of turning ambiguous technical capabilities into shipped products and measurable user or business outcomes.
  • Strong working knowledge of the modern model-development lifecycle, including data strategy, fine-tuning and preference optimization, evaluation, inference, experimentation, and production monitoring.
  • The technical judgment to reason with ML scientists and engineers about training objectives, reward design, data quality, model selection, scaling, latency, and serving cost - without pretending to be the scientist in the room.
  • Strong product judgment about where proprietary models create durable differentiation versus where external models or conventional systems are the better choice.
  • Experience creating clarity across multiple teams: defining decision rights, sequencing dependencies, resolving disagreement, and maintaining speed in a high-ambiguity environment.
  • A high bar for evidence, safety, and trust. You know that aggregate model scores can hide consequential failures and that clinical AI requires explicit escalation, abstention, and human-review paths.
  • Excellent written and verbal communication, including the ability to explain a technical strategy to both research teams and company leadership.

BONUS POINTS IF...

  • You have directly worked on LLM post-training, reinforcement learning, preference optimization, distillation, model routing, or domain-adaptive pretraining.
  • You have managed a portfolio spanning frontier APIs, open-source models, and in-house models.
  • You have experience with human-feedback or expert-annotation systems, especially where feedback is sparse, subjective, or expensive.
  • You have shipped AI in healthcare, life sciences, or another safety-critical and regulated domain.
  • You have worked on products involving clinical documentation, medical reasoning, agentic workflows, or longitudinal context.
  • You have partnered with research teams on high-compute bets where staged evidence was required before scaling investment.

Similar open jobs

Jobs in San Francisco (11741)