Forward Deployed Data Scientist, Lantum

• Ractical ML and Deep Learning (theory, methodologies, model architectures, and practical know-how around supervised, self-supervised, semi-supervised, and unsupervised paradigms, and ecologically valid evaluation)
• The wider Python (3) data science stack and ecosystem (such as Pandas, NumPy, Jupyter notebooks, SciPy, FastAPI, Flask, Matplotlib, and similar)
• Core ML and DL frameworks (such as PyTorch (strongly preferred), Keras, TensorFlow, scikit-learn, and similar)
• Cloud compute, infrastructure, services, and deployment w.r.t. end-to-end data science (ideally AWS (such as S3, EC2, Lambda, ECR, ECS))
• SQL and relational DBs (such as Postgres, Aurora or Athena)
• No-SQL DBs (such as MongoDB)
• Data visualisation methods and tools (such as Matplotlib, Bokeh or Seaborn)
• Basic REST (and GraphQL API) design for ML models and apps
• CI/CD
• ML app development and deployment platforms (such as Streamlit, BentoML, Gradio or PyCaret)
• Is a true scientist at heart, with genuine passion for the scientific process, scientific integrity, evaluating hypotheses, running experiments, evidence- and data-driven decision making, and overall precision-oriented excellence
• Is an effective, natural translator of complex technical and analytical topics and concerns between diverse audiences, covering technical vs. no-technical, junior vs. senior, and internal vs. external audiences, stakeholders, and colleagues
• Is not paralysed or frustrated by complex real-world problems and their eventual data representations “out in the wild” which may be regarded as underspecified, opaque, fragmentary, and intractable scientifically
• Is able to make judicious choices around structured strategic prioritisation and execution rationale from multiple angles to align, synchronise, balance, and complement diverse audience and their competing requirements
• Has demonstrable exposure to the modern, end-to-end data science process, life cycle, and best practices, with an awareness (and appreciation) of the role of product design and scalable software engineering in particular
• Excels at bootstrapping hiqh-quality data resources from raw unlabelled and partially labelled resources
• Has run substantial end-to-end (“from-idea-to-full-production”) data science projects involving non-trivial requirements (concrete technical as well as more open-ended business ones), and operational reliability and antifragility, resulting in performant models deployable rapidly to full long-term production
• Has a genuine customer-first mindset and passion required to complete substantial deliverables and work streams to customers under strict operational delivery objectives and SLAs
• Is able to learn a new domain quickly (particularly candidates without prior exposure to healthcare), to re-frame real-world problems and requirements at the right level of abstraction and in an optimal order, and to map analytical topics to concrete, executable development goals and features for other teams
• Is a credible, inspirational, and energetic scientist who can mentor and guide other data scientists and engineers
• Knows what genuine user focus really is: obsession with the real-life impact of the final end result – our products – on the lives of our customers

Desirable

• You will work directly on next-generation predictive models for complex and challenging staff and resource scheduling tasks which bring profound, directly measurable benefits for individual healthcare professionals and entire organisations alike
• At the same time, you will directly contribute to and execute Lantum’s AI strategy for a wider product platform; and implement and optimise company-wide data science processes, pipelines, workflows, resources, artefacts, and infrastructure
• This is a lead or senior role, depending on experience
• Working closely with other data scientists, data engineers, product teams, software engineering teams, and operations teams, the Forward-deployed Data Scientist will
• Implement and optimise production-ready, SLA’able models for multiple task domains such as complex staff and resource scheduling
• Interact directly with domain experts, healthcare practitioners, customers, and end users to design and deliver end-to-end data science work streams
• Support client onboarding teams in the field to capture client requirements, information architectures, data structures, data streams, key model requirements, acceptance criteria, and related data resources; and to measure, track, and optimise model and product performance
• Be a key contributor to day-to-day data science work at Lantum in a new team for end-to-end AI development and full-stack data science
• Contribute to Lantum’s foundational data infrastructure, data architectures, and production-grade data processes, workflows, and pipelines
• Bridge multidisciplinary teams covering both technical and non-technical audiences to enable rapid hypothesis testing and to promote effective data science practices (resources, people, and culture)
• Maintain fundamental scientific integrity across all AI and data science work through objective, measurable data-driven and evidence-based decision making and execution, and related policies and processes
• Provide scientific advice, both internally across the company and externally to clients