CV

Senior Data Scientist (Level III)

Idoven · Remote (Madrid, Spain)

• Led early detection research for cardiac amyloidosis on ~21K ECGs from ~3K patients; compared 9 deep learning architectures with Bayesian optimization, achieving AUC 0.88; co-authored publication in Heart Rhythm (2026)
• Tuned model operating thresholds and calibration for regulatory performance targets; extracted clinical validation metrics across patient subgroups; prepared SaMD technical documentation for CE mark submission
• Built reproducible data pipelines to ingest and clean multi-site clinical data from European and US hospitals, achieving 4.4x processing speedup; designed data collection forms for AstraZeneca partnership

Data Scientist

Anumana · Boston, MA

• Developed deep learning algorithm for early cardiac amyloidosis detection in partnership with Pfizer; defined target population, clinical indication for use, and statistical validation protocol contributing to FDA Breakthrough Device Designation
• Curated patient cohorts from 10M+ EHR records across multiple US hospitals (Spark SQL, Kedro); trained CNN and Transformer models on GPU clusters for coronary disease risk stratification with Novartis; published in eClinicalMedicine (The Lancet, 2023)
• Built NLP pipeline (BERT) to extract disease markers, symptoms, and lab results from unstructured Mayo Clinic patient notes; trained 3 gradient boosting models (AL, ATTR, combined amyloidosis) on the extracted clinical features achieving AUC 0.9; deployed in shadow mode alongside an ongoing Mayo Clinic amyloidosis clinical trial
• Rebuilt flagship LVEF detection model from scratch — new cohorts (500K patients), updated architectures, and retraining — then built a lifetime Markov model projecting QALY gains and cost savings for payer adoption

Machine Learning Instructor

United Nations International School & Inspirit AI · New York, NY

• Taught machine learning and deep learning to 200+ students at Inspirit AI (online ~30 students/cohort) and UNIS High School (UN International School, NYC); delivered content, coding sessions, and assignments across supervised/unsupervised learning, neural networks, computer vision, NLP, and ML ethics
• Built end-to-end research projects 1-on-1 with each student — EEG signal analysis, Parkinson's disease detection, skin cancer classification — from idea framing to working code to presentable result

Data Scientist Intern

CCC Intelligent Solutions · Chicago, IL

• Built deep learning architecture using RNN and pre-trained MobileNet for vehicle damage classification on 270K images; 5-point sensitivity improvement
• Evaluated telematics data inclusion in production models; presented integration recommendations to senior data science leadership

Deep Learning Graduate Researcher

University of Chicago SAND Lab · Chicago, IL

Prof. Heather Zheng

• Developed invisible adversarial ML patches for jamming facial recognition cameras; translated digital FGSM attacks to physical domain on robust body pose models
• Trained PixelCNN generative model to generate synthetic biometric sensor maps for testing impersonation attacks on EMS authentication devices; contributed to CHI 2021 publication

Computer Vision and Optimization Researcher

University of Chicago · Chicago, IL

Prof. Tingran Gao

• Developed non-convex optimization algorithms extending total variation methods to 3D graph domains; applied to filtering patterns from marine biology specimens (Bivalvia shells)