Senior Data Scientist · Healthcare

Baurjan Safi Clinical analytics, causal outcomes & applied AI

I work with claims and clinical data to help healthcare organizations decide who to reach, what works and what it is worth. My work spans risk models, intervention and ROI studies, and scaling AI-driven clinical solutions for high-growth startups. Before data science I spent 15 years in corporate finance and general management, so I build analyses around business decisions.

Get in touch Download resume (PDF)
Portrait of Baurjan Safi

Selected impact

Results from recent roles in value-based care and healthcare navigation.

**%higher engagement success from a clinically informed member stratification algorithm
**%less manual evaluation time with AI-powered clinical trial preselection
~7 minper person for clinical audit data preparation with an AI-assisted tool, down from hours
8+ yrsof healthcare data science, plus 8 years of corporate financial management

Experience

I have contributed to five startups, both as an individual contributor and as a team manager.

01/2026 – present
San Francisco, CA

Senior Data Scientist

Medzown
  • Engineered a clinically informed member stratification algorithm for targeted outreach, raising engagement success by **%.
  • Deployed AI-powered scaling tools for clinical trial preselection, cutting manual evaluation time by **%.
  • Led the build-out of the core analytics infrastructure, which delivers ROI, operational and clinical performance reporting.
Member stratificationGenerative AIAnalytics infrastructureROI analytics

Want to learn more about this work? Reach out.

09/2022 – 12/2025
San Francisco, CA

Senior Manager, Clinical Intelligence & Analytics

Accolade, Inc. (acquired by Transcarent)
  • Validated the clinical risk model.
  • Set the propensity score matching methodology for outcomes analysis.
  • Completed three comprehensive intervention analyses.
  • Developed the logic for six clinical performance guarantees.
  • Built an AI-assisted tool that cut clinical audit data preparation from hours to about 7 minutes per person.
Causal inferencePropensity matchingModel validationPerformance guarantees
04/2018 – 09/2022
Reston, VA

Senior Data Scientist

Somatus, Inc.
  • Developed and maintained an ML model that predicts the risk of inpatient hospitalization.
  • Trained a neural network to estimate the risk of developing a set of chronic conditions.
  • Recreated the logic of the CMS risk adjustment model for the finance team.
  • Created and formalized the logic for clinical metrics and utilization analysis.
Predictive modelingDeep learningCMS risk adjustmentUtilization
04/2018 – 12/2019
Washington, DC

Instructor Associate, Data Science

General Assembly
  • Taught in the Booz Allen Data Science 5K challenge (DC campus, Waves 2–7), about 120 students in total.
  • Taught regression, supervised and unsupervised ML, PostgreSQL, ERDs and Tableau. Average student rating: 4.5 out of 5.
Earlier career: finance & general management, 1994 – 2014
  • Regional Project Manager, KTIB Holding Company (2013–2014)
  • General Director, Petrodata Kazakhstan (2005–2013): operator of Kazakhstan’s National Oil & Gas Data Repository and a JV with Halliburton
  • General Director, MaxCom Ltd (2009–2013)
  • Finance Director / Director of Finance: Sodexho-KazMunaiGaz-Service, FoodMaster JSC (2001–2004)
  • Financial Analyst, Deputy Head of Planning, ACCEPT JSC (1999–2001)
  • Earlier roles as an administrative coordinator, an English–Russian interpreter and an English instructor

Skills

Grouped by the problems I solve.

Causal inference & outcomes

Measuring whether a program worked and what it was worth.

Propensity score matchingIntervention analysisROIPerformance guarantees

Predictive modeling & ML

Risk stratification and clinical prediction.

scikit-learnTensorFlowPyTorchModel validation

Applied AI

Putting AI into clinical and operational workflows, with people reviewing the output.

Generative AIAI-assisted reviewWorkflow automation

Data & engineering

Working at scale with large datasets.

SQLPythonpandas / NumPySpark (SQL, ML)PostgresMS SQLMySQL

Healthcare domain

Value-based care, member engagement and clinical quality.

Claims & clinical dataCMS risk adjustmentUtilizationClinical metrics

Communication & BI

Turning analyses into decisions for executives, clinicians and finance teams.

TableauPower BIQlikFinancial modelingPrice/cost modelingProject managementMentoring

Projects

Notebooks from my General Assembly Data Science Immersive (2017). Each one is an HTML version of a Jupyter notebook and links to its GitHub repository.

US Health Insurance Analytics

ACA marketplace premiums by state from 2014 to 2016, including smoker vs non-smoker rates.

Fake News: From Definition to Identification

NLP text classification capstone.

West Nile Virus in Chicago

Predicting outbreaks from weather, trap and spray data.

Global Terrorism: A Bayesian Approach

Bayesian inference on terrorism trends, including their links to oil prices.

Data Scientist Salaries in 21 US Cities

Web scraping and salary classification.

Iowa Liquor Sales Forecasting

Regression-based sales forecasting.

SAT Scores Across the USA

Exploratory data analysis and visualization.

Education

  • Data Science Immersive, General Assembly2017
  • Certificate in Corporate Finance, Georgetown University2016
  • MBA, KIMEP University1998
  • Diploma in Physics (6-year program, equivalent to an MSc), Kazakh National State University1996

Awards & service

  • 3rd place, Deloitte hackathon (medical insurance coverage)2017
  • 1st place, Pitch Contest, Association for Computing Machinery2017
  • Honorary Vice-Consul of Sweden in the Republic of Kazakhstan2011 – 2015

Let’s talk

I'm open to senior and staff data science roles in healthcare analytics, outcomes research and applied clinical AI. Coffee is on me.

baurjan.safi@gmail.com GitHub Medium Resume (PDF)