Data Scientist

Junction•New York, NY

1d•Remote

About The Position

Healthcare is in crisis and the people behind the results deserve better. With more and more data coming from wearables, lab tests, and patient–doctor interactions, we’re entering an era where data is abundant. Junction is building the infrastructure layer for diagnostic healthcare, making patient data accessible, actionable, and automated across labs and devices. Our mission is simple but ambitious: use health data to unlock unprecedented insight into human health and disease. If you're passionate about how technology can supercharge healthcare, you’ll fit right in. Backed by Creandum, Point Nine, 20VC, YC, and leading angels, we’re working to solve one of the biggest challenges of our time: making healthcare personalized, proactive, and affordable. We’re already connecting millions and scaling fast. Junction sits in the flow of high-value diagnostics and clinical data. As the company grows, our advantage moves beyond just having data to having the ability to turn it into reliable intelligence improving product decisions, customer outcomes, and the performance of the business. Some of that work exists today, but it is not yet owned as a coherent function. Models get built. Analyses get done. Experiments answer local questions. But we need someone who can define the broader scientific and analytical system: what we should measure, what methods we trust, where modeling creates real leverage, and how that work translates into products and decisions that hold up outside a demo. We’re hiring our first Data Scientist to take ownership of, and establish that standard. This role will lead Junction’s most important modeling, experimentation, and evaluation work. You’ll partner closely with data, product engineering and leadership teams to drive the analytical roadmap by which Junction can leverage differentiated value from data.

Requirements

Strong track record of leading high-stakes analytical work that influenced product, operational, or business decisions
Deep foundation in statistical inference, experimental design, observational analysis, and model evaluation
Strong Python and/or R skills, with experience working on large, messy real-world datasets
Experience building predictive or decision-support models in production or near-production environments
Experience partnering closely with engineering to move work from analysis or prototype into deployed systems
Ability to operate at both strategic and hands-on levels: defining the roadmap while also getting into the details when needed
Strong communication and stakeholder management skills; able to explain methods, findings, and tradeoffs to executives as well as technical peers
Comfort operating in a startup environment with ambiguity, limited structure, and high ownership

Nice To Haves

Experience designing, executing, and publishing research studies
Experience with HIPAA, PHI, or other regulatory clinical frameworks
Deep familiarity with modern data tooling and production workflows across warehouses, orchestration, and transformation layers
Experience developing, deploying, and designing evaluation frameworks for LLM or AI-powered features in customer-facing products
Expertise directly working with healthcare, diagnostics, lab data, wearable data, and other clinical data
Experience applying causal inference methods, such as diff-in-diff, propensity scoring, or instrumental variables in practice

Responsibilities

Own the research and modeling work underlying Junction’s highest-priority data science opportunities across diagnostics, clinical workflows, and AI-enabled product features
Define rigorous frameworks for measurement, experimentation, and causal evaluation so we can distinguish signal from noise and make decisions we can defend
Lead development of predictive models, segmentation approaches, risk or routing logic, and other statistical systems that directly inform product and business strategy
Build the analytical foundation behind customer-facing features — from model development through to validation and performance tracking
Partner with engineering and data engineering to ensure models and analytical systems can be put in production, are reliable, and useful in real workflows
Establish how Junction evaluates data-driven and AI-enabled features, including methodology, quality thresholds, monitoring, and performance review
Communicate complex technical findings clearly to technical and non-technical stakeholders, including tradeoffs, limitations, and implications for action