Adi Mendelowitz

Senior Data Scientist → ML Researcher

Senior data scientist with six years across medical AI and energy systems. At GE Healthcare I built NLP pipelines and MLOps infrastructure for clinical teams. At SolarEdge I led fleet data — monitoring, anomaly detection, and root cause attribution across millions of devices.

The work has always pulled toward the research end. These days that means building computer vision architectures from scratch — ResNet, ViT, Mask R-CNN, YOLOv8, SimCLR — and benchmarking them against published results. Most recently: lesion segmentation on ISIC 2018 medical imaging data, object detection on PCB defect datasets, and model compression down to 65x with measurable inference speedup.

Mathematics undergraduate. The code is on GitHub.

Download CV GitHub LinkedIn
Independent Research

Computer Vision & Deep Learning

Architecture implementations
PyTorch · torchvision · timm · CIFAR-10
ResNet-18 and Vision Transformer (ViT-Tiny) built from scratch in PyTorch, including multi-head self-attention and transformer encoder blocks. SimCLR contrastive pretraining reproduced from the original paper.
ResNet-18 93.4% ViT-Tiny 86.7% SimCLR 68.2% linear eval
Medical image segmentation & classification
Mask R-CNN · EfficientNet-B0 · ISIC 2018 · Kaggle T4
Mask R-CNN for skin lesion segmentation on ISIC 2018 Task 1. EfficientNet-B0 for 7-class disease classification on HAM10000. Winner benchmark included for reference.
Jaccard 0.782 Balanced acc. 0.746 Winner benchmark 0.802
Object detection
YOLOv8 · PCB defect dataset · Kaggle T4
YOLOv8n fine-tuned on a PCB defect detection dataset using a pretrained backbone with domain-specific augmentation. All six defect classes above 0.984 AP@0.5.
mAP@0.5 0.990 6 defect classes
Model compression
PyTorch · INT8 quantization · Knowledge distillation
Static INT8 quantization and knowledge distillation benchmarked against FP32 baseline. Latency and throughput measured across all model variants.
3.95x size reduction 1.52x inference speedup 65x compression
Professional Experience
SolarEdge Technologies Mar 2022 – Aug 2025
Senior Data Science Engineer · Fleet Data Team Lead · Data Scientist
  • Built telemetry analysis pipelines monitoring millions of inverters and battery units across heterogeneous device fleets at 24-hour latency; drove anomaly detection and root cause attribution at scale.
  • Led failure and survival analysis across large heterogeneous device fleets; improved preemptive failure detection by 25% and reduced maintenance costs by 20%.
  • Automated warranty cost projection models, reducing processing time by 90% and direct costs by 15%.
  • Built a Python/Streamlit SQL interface so non-analyst managers could query fleet data without analyst support.
  • Built multilingual translation pipelines across three to four languages for thousands of recorded customer-support conversations; validated output quality through iterative review with Tier 1/2 domain experts.
  • Mentored a team of 4 data scientists and engineers; established Agile practices and code quality standards.
GE Healthcare – Life Care Solutions Nov 2019 – Jan 2022
NLP Data Operations Engineer · Medical AI · MLOps
  • Built NLP pipelines processing clinical speech and text from hospital teams across multiple sites; curated a 10,000+ utterance ground truth dataset to train and evaluate a conversational interface for medical professionals.
  • Built MLOps pipelines for continuous model retraining with a production/shadow deployment pattern; automated performance diagnostics and drift detection determined whether new models replaced the live version.
  • Automated annotation pipelines by routing utterances to cluster-specific tagging workflows based on intent domain; replaced a manual multi-step process with backend automation.
  • Supervised a team of 2–4 engineers and 10+ remote clinical domain experts across annotation, evaluation, and model development.
Education
B.Sc. Mathematics
The Open University of Israel
Completed 2021

Part-time while employed full-time.

Contact
adimendelowitz@gmail.com LinkedIn GitHub