You will get Build an explainable baseline fast (Mini-Model + SHAP)
Rising Talent
Rising Talent
Project details
Also available for custom scopes. Share a schema + 1–2% sample and I’ll confirm price/timeline within 24 hours.
Need clean data first? See Healthcare Data Check
I deliver an explainable, business-ready baseline fast, and I make it easy for non-technical stakeholders to trust and act on it. My focus is clarity and rigor: a clean, reproducible pipeline (repo or notebook, README, and optional MLflow tracking) plus a concise 1–2 page executive brief that ties model metrics to your business goals.
Using SHAP, I provide global and local explanations so you can see which features drive predictions and why specific cases behave the way they do. If helpful, I add fairness and slice checks and threshold recommendations to balance impact versus risk. I have experience in healthcare and analytics, so I am comfortable with data quality, leakage traps, and privacy constraints. You will get pragmatic guidance on what to do next, whether productionizing, iterating on features, or spinning up a lightweight Streamlit demo.
Need clean data first? See Healthcare Data Check
I deliver an explainable, business-ready baseline fast, and I make it easy for non-technical stakeholders to trust and act on it. My focus is clarity and rigor: a clean, reproducible pipeline (repo or notebook, README, and optional MLflow tracking) plus a concise 1–2 page executive brief that ties model metrics to your business goals.
Using SHAP, I provide global and local explanations so you can see which features drive predictions and why specific cases behave the way they do. If helpful, I add fairness and slice checks and threshold recommendations to balance impact versus risk. I have experience in healthcare and analytics, so I am comfortable with data quality, leakage traps, and privacy constraints. You will get pragmatic guidance on what to do next, whether productionizing, iterating on features, or spinning up a lightweight Streamlit demo.
Machine Learning Tools
pandas, Python, Python Scikit-Learn, scikit-learn, XGBoostWhat's included
| Service Tiers |
Starter
$595
|
Standard
$995
|
Advanced
$1,995
|
|---|---|---|---|
| Delivery Time | 4 days | 6 days | 9 days |
Number of Revisions | 2 | 2 | 2 |
Number of Model Variations | 1 | 3 | 3 |
Number of Scenarios | 1 | 2 | 3 |
Number of Graphs/Charts | 4 | 6 | 8 |
Model Validation/Testing | |||
Model Documentation | |||
Data Source Connectivity | - | - | |
Source Code |
About Adriana
Healthcare Data Scientist | ML Pipelines | EHR Analytics | HIPAA
Constitucion, Chile - 3:08 pm local time
With 6+ years in US healthcare data science, I specialize in two things that tend to go together: getting messy clinical data into shape, and taking ML from notebook to production in a way that is reproducible, explainable, and compliance-aware.
Some past results:
— ML models forecasting medical absences → ~$100K in operational savings
— SQL audit uncovering 700 uncancelled appointments for deceased patients → est. $70K–$140K in recoverable savings
— Ranking models improving performance metrics for ~200 clinicians by an average of 10%
— 4 ML prototypes productionized with Docker, CI/CD, and on-premise HIPAA-compliant deployment
I work with EHR, claims, and population health data, and I know the standards well enough to catch what a generalist misses. I am familiar with FHIR, SNOMED CT, LOINC, ICD coding inconsistencies and interoperability gaps.
What I typically deliver:
— Data quality audits with prioritized findings and risk notes
— Modular, tested Python repos with CI/CD — not notebooks handed off and forgotten
— Baseline ML with leakage-safe splits, calibrated metrics (F1, PR-AUC), and SHAP explainability
— End-to-end deployment: Docker · Prefect · MLflow · FastAPI · AWS/GCP
Scope stays clear, handovers are practical, and I keep clinical and non-technical stakeholders in the loop throughout.
Stack:
Python, SQL, R, PostgreSQL, Pandas, Scikit-learn, XGBoost, SHAP, MLflow, Prefect, Docker, FastAPI, GitHub Actions, AWS, GCP, ETL, EHR, FHIR, HIPAA, Data Quality, Machine Learning, Statistical Modeling, EDA, NLP
Steps for completing your project
After purchasing the project, send requirements so Adriana can start the project.
Delivery time starts when Adriana receives requirements from you.
Adriana works on your project following the steps below.
Revisions may occur after the delivery date.
Kickoff & Data Check
Review goals/metric, validate sample data/schema, confirm leakage rules and privacy.
Baseline Modeling & Metrics
Clean/split data, train 1–3 baseline models, report core metrics (F1/PR-AUC/MAE).