Full-Time
Product Manager
AI/ML, Evals
Abridge
501-1,000 employees
AI-powered platform for clinical docs
Compensation Overview
$250k - $290k/yr
San Francisco, CA, USA
Hybrid
Hybrid work in the San Francisco office.
 Product (1) |
|---|
Get referred to Abridge
See people who can refer or advise you
- At least 5 years of product management experience with significant ownership of machine-learning-powered products or platform systems.
- Deep understanding of measuring and improving model quality, including evaluation frameworks, annotation pipelines, and benchmark design.
- Strong technical fluency across machine learning, data pipelines, and distributed systems.
- Experience working closely with machine learning researchers and engineers to drive impact in production.
- Ability to balance long-term architectural investments with near-term quality improvements.
- Strong communication skills and ability to translate complex technical concepts into clear decisions and narratives.
- Track record of delivering high-quality products in domains where accuracy, reliability, and trust are paramount.
- Drive product strategy and execution for the evaluations platform, own the roadmap across the evaluation lifecycle, and own outcomes against it.
- Build shared measurement infrastructure that enables teams to define quality, run experiments, compare models, and monitor production.
- Own standards for large language model judges, rule-based evaluators, human annotation, and online monitoring, including deciding where each approach belongs.
- Create a repeatable process for teams to evaluate a new frontier model within days of release and establish a defensible framework for choosing between a post-trained specialty model and a prompted frontier model.
- Keep the evaluation system model-agnostic so it remains a neutral referee.
- Define the evaluation gates from early build through general availability and steady-state monitoring, determine which gates are hard versus advisory, and establish ownership of non-negotiable critical-error-rate floors across pods without formal authority.
- Work cross-functionally with Data Engineering on the de-identification pipeline, Data Science on improving judge quality with less human annotation, Clinical Science on flagged production cases, and the agent platform team as workflows become agentic.
- Experience building evaluation platforms, machine learning observability systems, or quality measurement pipelines.
- Experience working in clinical, healthcare, or regulated environments with a high bar for accuracy and compliance.
- Experience working on specialty-specific or domain-specific model adaptations.
- Experience working on personalization systems, context ingestion frameworks, or ambient intelligence products.
- Experience shipping large-scale machine learning products with human-in-the-loop workflows.
Abridge builds an AI-powered platform that turns recorded medical conversations into structured clinical documents. Clinicians record patient visits, and the system processes the audio to create organized notes that capture essential data and the clinical narrative. The product is delivered through a subscription model to healthcare providers, payers, and pharmaceutical companies, aiming to save time on documentation and support better patient care. Abridge differentiates itself by focusing on converting conversations into actionable, structured documentation that integrates into healthcare workflows for multiple stakeholders, rather than offering generic transcription. The overall goal is to improve patient care and provider efficiency by streamlining clinical documentation and freeing clinicians to focus more on patients.
Company Size
501-1,000
Company Stage
Acquired
Total Funding
$757.5M
Headquarters
Pittsburgh, Pennsylvania
Founded
2018
Get referred to Abridge
See people who can refer or advise you
Simplify's Take
What believers are saying
- Abridge says 300-plus health systems support over 100 million conversations annually.
- Northwestern Medicine rolled out Abridge enterprise-wide in June 2026, validating large-system adoption.
- June 2026 partnerships with Nvidia, Lilly, and Artisight broaden product distribution and data access.
What critics are saying
- Sutter Health and MemorialCare face April 2026 class actions over Abridge consent practices.
- Epic and Microsoft keep bundling ambient documentation into existing EHR and hospital software.
- Clinical-workflow expansion multiplies privacy, billing, and audit failures across 300-plus health systems.
What makes Abridge unique
- Abridge, on June 11, 2026, spans pre-visit, in-visit, and post-visit workflows.
- NVIDIA and Abridge are co-developing a clinical-conversations foundation model on Blackwell infrastructure.
- Eli Lilly invested in June 2026, linking Abridge directly to trial-screening workflows.
Help us improve and share your feedback! Did you find this helpful?
Benefits
Medical insurance
Dental Insurance
Vision Insurance
Unlimited Paid Time Off
Equity
Flexible Spending (FSA) & Health Savings (HSA) Accounts
Learning and Development budget
401k Plan
Paid Parental Leave
Flexible working hours
Home Office Budget
Sabbatical Leave
Growth & Insights and Company News
6 month growth
↑ 0%1 year growth
↑ 1%2 year growth
↑ 4%Top Pharma news: July 2026. Posts You Might Like What were the top pharma news stories as of July 2026? July's pharmaceutical landscape was shaped by continued U.S. drug pricing reform, a breakthrough in cancer treatment innovation, and expanding investment in artificial intelligence across healthcare. The Trump administration moved to strengthen Medicare drug pricing regulations, China approved the world's first CAR-T therapy for solid tumors, and leading healthcare organizations accelerated AI adoption to improve drug development, clinical decision-making, and specialty pharmacy services. The Fullintel Hub keeps stakeholders informed as these developments unfold. June 2026 Pharma news breakdown. What does the proposed CMS rule signal about the administration's drug pricing strategy? The Trump administration proposed a comprehensive rule through the Centers for Medicare and Medicaid Services to codify federal Medicare drug price negotiations permanently. This measure closes a major regulatory loophole that previously allowed pharmaceutical companies to avoid mandated price caps simply by adding active ingredients or slightly reformulating existing single-source medications. What it means: * The proposal strengthens the long-term stability of Medicare drug price negotiations by reducing opportunities for legal or regulatory challenges. * Drug manufacturers may face increased pressure to align pricing strategies with negotiated Medicare prices, closing off traditional product-hopping workarounds. * The move reinforces drug affordability as a central healthcare policy priority while increasing scrutiny of pharmaceutical pricing practices. Social media sentiment was generally positive, with "Love" and "Wow" reactions reflecting support for closing corporate loopholes and reducing out-of-pocket prescription costs for seniors. Limited negative sentiment likely reflected skepticism over whether drugmakers might find new regulatory workarounds or reduce overall investment in drug discovery. Why is CARsgen's approval considered a milestone? China approved CARsgen Therapeutics' satricabtagene autoleucel, making it the world's first CAR-T therapy approved for treating solid tumors. The therapy targets Claudin18.2-positive gastric and gastroesophageal junction cancers, representing a major scientific breakthrough after years of limited success applying CAR-T technology beyond blood cancers. What it means: * The approval establishes proof of concept for CAR-T therapies in solid tumors, addressing one of oncology's biggest unmet challenges. * China continues strengthening its position as a global leader in advanced cell and gene therapy innovation, raising questions about when Western regulatory bodies like the U.S. FDA will follow. * The decision is expected to accelerate investment in next-generation CAR-T platforms targeting additional solid tumor indications, opening doors for broader commercial applications in hard-to-treat cancers. Social media sentiment was generally positive, with dominant "Love" and "Wow" reactions reflecting support for lower out-of-pocket drug costs and innovative medicine. The pulse: AI in healthcare. 1. Nvidia, Abridge collaborate to develop healthcare-specific AI model. Nvidia and healthcare AI company Abridge announced a collaboration to develop healthcare-specific foundation models designed to improve clinical documentation, physician workflows, and medical decision support. The partnership combines Nvidia's AI infrastructure with Abridge's expertise in clinical language models to expand generative AI adoption across healthcare systems. * Coverage Sentiment: Positive. Analysts view the partnership as another major step toward specialized AI models built specifically for healthcare applications. * Dominant Social Sentiment: "Wow" - Healthcare technology communities praised the collaboration for addressing clinician burnout while improving documentation accuracy. * News Source: Fierce Healthcare 2. Cigna invests $100M in new ai-powered specialty pharmacy program. Cigna announced a $100M investment to launch a new AI-powered specialty pharmacy program designed to improve patient support, streamline prior authorization processes, and optimize medication management for complex diseases. * Coverage Sentiment: Positive. Coverage emphasized operational efficiency and improved patient engagement through AI-enabled pharmacy services. * Dominant Social Sentiment: "Love" - Discussion focused on AI's potential to simplify specialty pharmacy experiences while reducing administrative burdens. * News Source: Reuters 3. AstraZeneca CEO says AI is reshaping drug development and helping boost the odds of success. AstraZeneca CEO Pascal Soriot stated that artificial intelligence is significantly improving the probability of success throughout the drug discovery and development process. AI is increasingly being integrated into target identification, clinical trial design, and molecule optimization to accelerate research while reducing development costs. * Coverage Sentiment: Positive. Industry observers increasingly view AI as a competitive advantage for pharmaceutical R&D. * Dominant Social Sentiment: "Wow" - Conversations centered on AI's growing role in transforming pharmaceutical innovation and improving research productivity. * News Source: CNBC Key pharmaceutical industry takeaways for July 2026. What do these developments signal for the health and life sciences industry? * Medicare drug pricing reform continues to evolve, with policymakers focusing on strengthening the regulatory framework supporting negotiated drug prices. * Cell therapy innovation has entered a new phase as CAR-T technology expands beyond blood cancers into solid tumor treatment. * AI investment is accelerating across the healthcare ecosystem, extending beyond research into clinical operations, pharmacy services, and physician workflows. * Specialized healthcare AI models are emerging as the preferred strategy for improving clinical accuracy and operational efficiency. * Global competition in pharmaceutical innovation continues to intensify, with China demonstrating increasing leadership in advanced oncology therapies. About Fullintel's pharmaceutical news Coverage. The Fullintel Hub delivers daily pharmaceutical news analysis and insights to support PR and communications teams. Its human analysts track key developments across traditional media, social platforms, and industry publications. Its award-winning methodology, including its proprietary Media Impact Score and Trust Score, helps organizations refine messaging and benchmark against competitors. About Fullintel Strategic Media Analysis. Fullintel's Strategic Media Analysis measures the true impact of PR and communications efforts for pharmaceutical and healthcare organizations. Angela Dwyer, VP of Insights, leads its pharmaceutical analysis practice. Her original trust factor research has been presented at PRSA ICON and IPR conferences, and she has been named a PRNEWS Top Women Award recipient and 2025 AMEC Award winner. Angela Dwyer. Angela is VP of Insights at Fullintel - a media intelligence company that specializes in news monitoring and analysis. She has worked in media measurement for 15 years, helping brands improve business results through data-driven, actionable insights. From public relations agencies like Lippe Taylor to media research firms like PRIME Research, she has consulted across industries, particularly healthcare and pharmaceuticals. She has presented and published several award-winning research papers about news content that drives recall, engagement, and brand trust. Her "Trust in Pharma" research outlines how biopharma brands can build and sustain trust. She contributes knowledge at the intersection of academia and practice as director of the International Public Relations Measurement Commission and as a member of the International Public Relations Research Conference Board. Her contributions have been recognized with multiple industry awards, including PRNEWS People of the Year (Data & Measurement Game Changer), PRNEWS Top Women (Industry Champions), and AMEC Rising Star for innovation in communication measurement. Learn more about Angela Dwyer's expertise, industry recognition, and thought leadership.
Inventurus Knowledge Solutions Limited announced that its wholly-owned subsidiary, Inventurus Knowledge Solutions, Inc., has sold its stake in Abridge AI Inc. for $10 million. The subsidiary entered into a Stock Purchase Agreement to sell 54,055 Series A-1 Preferred Shares to Alpha Square Group Fund IV, LP. The transaction, dated 5 August 2026, represents a partial divestment by the subsidiary. The company confirmed that the purchaser is not related to the promoter, promoter group, or group companies of Inventurus Knowledge Solutions Limited. The deal was conducted on an arms-length basis and is not considered a related party transaction.
Abridge, an AI medical scribing company, has acquired Altrina, a startup that builds AI agents to automate multi-step workflows across various industries, including healthcare. Altrina was founded in 2025 and backed by Y Combinator. The acquisition brings Altrina's co-founders Mo Nasir and Harvey Hu, plus founding engineer Karthik Prasad, to Abridge. The team has experience building browser agents, automating workflows across electronic health records, and developing production infrastructure. Abridge plans to use this expertise to develop technology reducing administrative burdens for clinicians. The company recently expanded its AI scribing platform to include enhanced support before, during, and after patient visits, featuring pre-visit notes, ambient clinical documentation, and follow-up summaries. Eli Lilly recently made a strategic investment in Abridge to support research.
Abridge, a clinical-AI startup, is launching its Clinician Intelligence Platform in partnership with Nvidia, CEO Dr Shiv Rao announced. The platform transforms clinician-patient conversations into computable workflows that assist throughout the entire visit process. Rao described the technology as an operating system for vertical AI companies targeting regulated domains, emphasising the importance of evaluations in the healthcare sector. The launch represents a significant competitive shift in the clinical AI space.
Abridge launches a patient-centered AI platform to connect care, payment, and clinical evidence. The ambient AI company is moving well beyond note-taking, with NVIDIA building a clinical foundation model to power its next generation of tools June 17, 2026 Abridge built its name on one thing: using AI to listen to doctor-patient conversations and turn them into clinical notes automatically. That was already useful enough to land the company in more than 300 health systems. Now the company wants to do much more. At a keynote event in New York on June 11, Abridge announced what it's calling a patient-centered clinician intelligence platform. The idea is to bring AI into every stage of a clinical encounter, from preparation before the visit to billing and claims after it, and to connect health systems with payers and life sciences organizations through those same clinical workflows. The company also announced a collaboration with NVIDIA to build what both companies describe as the first foundation model built specifically for clinical conversations. Northwestern Medicine, one of Chicago's largest academic health systems, was named as a new enterprise-wide customer. How does it work? Abridge organizes its platform around three stages of the clinical encounter: * Before the visit: The system generates pre-charted notes and patient summaries tailored to the care setting. For hospital rounds, it pulls context from emergency department notes, nursing assessments, labs, and imaging. For outpatient visits, it surfaces a draft history and flags relevant care gaps or chronic conditions before the clinician walks in. * During the visit: Abridge now goes beyond passive transcription. It can suggest discussion topics based on the patient summary and the ongoing conversation. Clinicians can query clinical evidence without leaving the platform. The speech recognition supports more than 28 languages across specialties. * After the visit: The platform generates clinical notes, flowsheets, billing codes, patient summaries, and order drafts. Clinicians can edit outputs using plain language before everything flows into the electronic health record. Abridge integrates with Epic, Oracle Health, and athenahealth. Abridge is also expanding to nurses. The system can capture nurse-patient conversations during inpatient care and turn them into structured EHR documentation, with relevant findings carried forward into the next clinician's pre-visit summary. On the payment side, Abridge is working on real-time claims workflows, aiming to align documentation and billing at the moment care is delivered rather than days or weeks later. The company is partnering with AHIMA, the health information management association, to validate the accuracy and auditability of its coding outputs. For life sciences, Abridge is exploring how clinical context captured during a conversation could help identify patients who may be eligible for clinical trials, including early screening for conditions like Alzheimer's disease where biomarkers can appear years before diagnosis. The NVIDIA collaboration centers on a new foundation model built on the Nemotron open model family, trained on NVIDIA Blackwell infrastructure using de-identified clinical data. Unlike generic large language models adapted for healthcare after the fact, this model is designed to learn clinical reasoning from the ground up across all three training stages. Why does it matter? Physician burnout is a well-documented crisis, and documentation load is one of the biggest drivers. Studies have consistently shown that clinicians spend more time on administrative tasks than on direct patient care. Ambient AI tools like Abridge address that directly, and the evidence from early adopters is starting to look compelling. Reid Health reported that after rolling out Abridge for nurses, its nursing vacancy rate dropped from 18% to 8.6% with no contract staff, and incidental overtime fell 70% on the teams using it. The hospital's chief nursing officer noted that retraining a single nurse costs close to $100,000, which puts the financial case in sharp relief. The payment angle is less developed but potentially more significant. Billions of dollars in care goes unrecognized every year because clinical complexity gets lost between the visit and the claim. If Abridge can ground billing codes and claims in a verified record of what was actually said and done, it could reduce the friction between providers and payers that currently slows down reimbursement and creates costly disputes. The foundation model collaboration with NVIDIA matters because most AI in healthcare today is built on general-purpose models fine-tuned for medical use. A model trained from the start on clinical conversations, across specialties and care settings, could produce more accurate and auditable outputs, which matters enormously in a high-stakes environment where errors have real consequences. The context. Abridge is not the only company in this space. Competitors including Nuance (owned by Microsoft), Suki, and Nabla all offer ambient documentation tools, and Epic has been building its own AI features into its EHR platform. The market is crowded and consolidating fast. What Abridge is trying to do with this announcement is reframe the competitive conversation. Instead of competing on transcription accuracy alone, the company is positioning itself as infrastructure for the entire clinical and financial workflow, which is a much harder category to displace. The clinical decision support additions are also notable. Abridge has signed content agreements with the American Diabetes Association, the American Academy of Family Physicians, the American Heart Association, the New England Journal of Medicine, JAMA, and the Journal of Clinical Oncology, among others. Clinicians using the decision support tools will also be able to claim continuing medical education credit. That kind of ecosystem building takes time to replicate. Abridge says it now supports more than 100 million conversations annually across health systems that collectively serve more than 250 million patients. Those numbers give the company a significant data advantage as it trains the new NVIDIA model, assuming appropriate de-identification and governance processes hold up to scrutiny. That scrutiny, from regulators, privacy advocates, and the health systems themselves, will be one of the more important tests the company faces as it scales.