Our Intelligence Suite transforms operating room (OR) performance, keeping teams in sync and workflows on track to maximise throughput. Simultaneously our computer vision and AI capabilities capture real-time data and detect surgical events – improving quality of data outputs. The result: ORs are optimised like never before – with predictive analytics and automated notifications ensuring patients and staff are in the right place at the right time. 

Once practitioners are in the OR, our Surgical Suite enables real-time remote access and creates a secure video record of every procedure; improving training, education, and collaboration. It is an intuitive asset which helps instil a culture of continuous learning, accelerates the adoption of cutting-edge medical devices, and enhances surgical performance across the entire global workforce – improving outcomes and saving lives. 

Proximie was commercialised in 2019 and is available in over 500 facilities globally. 

Check out our Founder and CEO Nadine’s Origins Story here: https://www.proximie.com/about-us/ 

Position Overview

As Proximie continues to turn every activity and event in the operating room into comprehensive, structured, and context-rich data that drives better insights and decision making, using your deep technical expertise and real-world experience you will be instrumental in developing machine learning solutions that improve clinical outcomes, drive productivity and support Proximie's continuing evolution into the unrivalled champion of the intelligent operating room.

You will work with multi-modal data (audio, vision, and language) to curate, consolidate, and augment retrospective and prospective datasets that fuel the development of advanced machine learning solutions. You will harness the power of Proximie's deep data lakes and apply cutting-edge generative techniques to solve real-world challenges for Proximie's customers. You will also lead the development of intelligent systems that can automatically detect and capture key events in the operating room, designing robust solutions that work across diverse and unpredictable data distributions, tackling challenges like identifying rare events buried in hours of surgical video.

This is a commercial team. The problems you work on come from hospitals rather than from a research agenda, and the models you build are judged on whether they change how surgery is run day to day. Our MLOps engineers own the platform that takes your work into production, which leaves you free to focus on building models that hold up in the real world.

This role demands creativity, precision, and a deep understanding of real-world machine learning at scale. If you are excited by complex problems with life-changing impact and want to build tech that operates in the most critical environments, we would love to hear from you.

Job Responsibilities

• Own the full model lifecycle, including but not limited to data curation, model implementation, GPU based training and validation.

• Design, train and validate machine learning models across vision, audio and language data.

• Curate, consolidate and augment retrospective and prospective datasets from Proximie's data lakes, and build the annotation workflows that support them.

• Lead the development of systems that automatically detect and capture key events in the operating room, including rare events buried in hours of surgical video.

• Build models that hold up across the range of operating room environments we deploy into worldwide.

• Set the evaluation bar for a model: define what good looks like clinically and commercially before training starts.

• Work with the MLOps engineers to get models into production and improve them on real-world performance data.

• Partner with product and commercial teams to shape what the Intelligence Suite does next.

• Document your work and share what you learn across the engineering team.

Requirements

Essential

• PhD in machine learning, computer science, data science or a related field. A Master's with strong industry experience will be considered.

• 4+ years in industry developing and deploying AI solutions that reached real users.

• ·Expertise developing, training and fine-tuning multimodal models on heterogeneous data distributions.

• Strong grounding in traditional ML, deep learning and generative AI across supervised, self-supervised and unsupervised approaches.

• Proficiency with deep learning frameworks such as PyTorch.

• Experience with LLM frameworks, for example vLLM, is highly desired.

• Experience optimising models using ONNX, TensorRT or similar frameworks.

• Experience with Docker and containerised ML workflows.

• Experience developing annotation workflows.

• Strong Python and software engineering skills.

• Able to explain complex technical concepts to clinical, commercial and executive audiences.

Desirable

• Experience analysing surgical multi-modal data is highly desired.

• Exposure to healthcare, medical devices or another regulated industry.