Clinical AI Engineer-Intern

CARDIO AI

STEM Internship Program

Clinical AI Engineer-Intern

Dublin, OH | San Francisco, CA | Remote-Friendly | www.cardioailive.com

Cardio AI is a clinical AI platform that helps clinicians catch and manage heart disease earlier. We run five risk calculators at once — covering heart attack, stroke, heart failure, sudden cardiac death, and irregular heartbeat — then guide prevention, treatment, and ongoing care. A dedicated tool for women catches six serious heart conditions that are routinely missed, including pregnancy-related heart failure and a type of heart attack that mainly affects younger women.

Our STEM Internship Program gives students and recent graduates the opportunity to work on real clinical AI problems alongside our engineers and clinical advisors. This is not a coffee-and-spreadsheet internship. You will write code that runs on real medical data and contributes to a platform designed to save lives.

Duration

10-12 weeks

Summer | Fall | Spring

Compensation

Unpaid

Academic credit eligible

Who

Undergrad | Graduate

Recent grads welcome

Clinical AI Engineer — Intern

Cardio AI STEM Internship | Unpaid · Credit Eligible | Remote / Dublin, OH / San Francisco, CA

Location

Remote (preferred). On-site option in Dublin, OH or San Francisco, CA.

Duration

10-12 weeks. Summer, Fall, or Spring cohort.

Compensation

Unpaid. Academic credit may be arranged through your institution.

Eligibility

Current undergraduate or graduate student, or recent graduate (within 12 months).

Fields

Computer Science, Biomedical Engineering, Data Science, Statistics, or related.

Apply

tonywell@cardioailive.com | Subject: Clinical AI Intern Application

As a Clinical AI Engineer Intern you will work on real machine learning models used in cardiovascular disease prediction. You will be embedded in our engineering team, contributing to one of our core AI workstreams — processing ECG waveforms, analyzing echocardiography images, or building structured risk prediction models. By the end of your internship you will have trained or evaluated a model on real clinical data and understand what it takes to move AI into a clinical setting.

You will be placed on one of three project tracks based on your background and interest:rack A — ECG anaveform Anay

•     Pre-process 12-lead ECG recordings: remove noise, filter signals, and segment waveforms into model-ready patches.

•     Help train and evaluate our ECG classification model on rhythm tasks — atrial fibrillation detection, STEMI identification, bundle branch block classification.

•     Run performance analyses by patient age, sex, and recording device to check consistency across groups.

•     Visualize which parts of the ECG the model focuses on for each arrhythmia — making the model explainable to clinicians.

Track B — Echocardiography Image Analysis

•     Process echocardiography image files — extract frames, resize, normalize, and organize multi-view heart ultrasound studies for model training.

•     Fine-tune our image recognition model for ejection fraction prediction on an external validation dataset.

•     Compare model predictions against cardiologist measurements and compute accuracy metrics.

•     Build data augmentation pipelines to improve how well the model generalizes across different hospitals and equipment.

Track C — Cardiovascular Risk Models

•     Work with structured patient data — lab results, vital signs, diagnoses, medications — to build risk prediction models for one of our five cardiovascular calculators.

•     Tune and evaluate a model predicting heart attack, stroke, heart failure, sudden cardiac death, or atrial fibrillation risk against published clinical benchmarks.

•     Compute performance metrics including AUC-ROC, calibration, sensitivity, and specificity.

•     Use SHAP to identify the most important features driving predictions and explain findings to a non-technical audience.

You are a strong candidate if you:

•     Have taken at least one machine learning or deep learning course and understand the basics: training loops, loss functions, overfitting, evaluation metrics.

•     Can write clean Python and are comfortable with at least one ML library: PyTorch, TensorFlow, or scikit-learn.

•     Are curious about healthcare and want to understand how AI models connect to real clinical decisions, not just benchmark scores.

•     Can take ownership of a well-scoped problem, ask good questions when stuck, and communicate clearly about what you find.

Nice to have — not required:

•     Exposure to medical data: ECG files, ultrasound images, electronic health records, or clinical research datasets.

•     Coursework in biomedical signal processing, computer vision, or time-series analysis.

•     Familiarity with NumPy, Pandas, and matplotlib for data exploration.

•     Comfort with Git and running experiments in Jupyter notebooks.

•     A real project on your resume — not a tutorial, a model trained or evaluated on actual clinical data with documented results.

•     Weekly mentorship from a senior engineer or clinical advisor who is invested in your growth, not just your output.

•     A practical understanding of the full clinical AI lifecycle: raw data, model training, validation, and the path to clinical deployment.

•     A letter of recommendation and professional reference upon successful completion.

•     Priority consideration for full-time roles as Cardio AI grows — we hire from our intern cohort first.