Machine Learning & Data Science Engineer at Tricura Insurance Group - a US-based healthcare startup specializing in risk management and claims innovation

Hire5•New York, NY

6h•Remote

About The Position

Tricura Insurance Group is a next-generation, tech-enabled insurance and risk-management company transforming how healthcare organizations protect, manage, and grow their operations. We’re a global remote-first team working to reshape the future of healthcare insurance. We combine deep industry expertise with cutting-edge technology to deliver smarter, faster, and more transparent insurance solutions across the continuum of care. If you're passionate about creating impactful solutions that empower those who care for others, we’d love to meet you. The title spans both disciplines because the work does. You'll analyze data and build models with sound methodology (data science), and you'll design how those models integrate into the product: architecture, deployment, cost, SLAs (ML engineering). You'll interface regularly with backend engineering on integration and with product/UX on how outputs are surfaced. The only thing not in scope is dashboards or ad-hoc business reporting. Analysis happens, but always in service of a modeling goal. Small, fully hands-on ML team with no middle management. Everyone carries their own project load. We have structured support: semiannual performance reviews, regular 1:1s, and a culture where everyone is accessible and willing to help. There's no ML platform team, no feature store, no curated datasets. You work directly with data across multiple systems. The flip side: everything you build ships to production with direct product impact. No research-only projects, no models on a shelf. Most of your time goes into figuring things out. A typical project starts with a business goal. From there you determine what data exists, where it lives, what's missing, and how to get it (internal databases, public sources, third-party providers). You define the target, build and validate the model, and deliver something the team can trust and deploy. Problems are loosely defined, data is imperfect, and dead ends happen, but nothing is wasted: if you're building it, the business needs it. For the right person, this is a growth multiplier. The breadth of problems and level of autonomy mean you won't repeat the same type of task for years. Your skills compound fast.

Requirements

Python and the data science / ML ecosystem
Applied ML: regression, classification, gradient boosting, time-series. End-to-end (build, tune, validate, deploy)
SQL fluency (Postgres, Snowflake, or similar). Comfortable with unfamiliar schemas
Cloud platforms (AWS preferred: S3, EC2, SageMaker)
Git
Project ownership, from problem definition through delivery
Self-management: own your priorities, time, and quality bar
Comfort with ambiguity. Figuring things out when the path isn't clear is the norm, not the exception
The instinct to question your own results before moving on
Clear communication: knowing when to say "I'm exploring" vs. "I'm delivering" vs. "I'm stuck."

Nice To Haves

Healthcare, insurance, or finance background
Underwriting or risk scoring experience
NLP (classical or LLM-based) as a complement to traditional modeling
AI-assisted coding tools (we use Claude Code a lot)

Responsibilities

Build and deploy predictive models for insurance pricing, risk assessment, and claims forecasting
Source, clean, and engineer features from internal, external, and third-party data. Expect significant time on data work
Validate rigorously: train/test splits, cross-validation, baselines, error analysis. Understand why your model works, not just that it runs
Deploy to production via APIs and cloud infrastructure. Monitor and retrain as needed
Communicate results to technical and non-technical stakeholders. Tie model outputs to business decisions