Adi Mendelowitz

Data Scientist · Applied Machine Learning

Data scientist with six years of applied machine learning in production, across medical AI and energy systems. At GE Healthcare I directed the design of clinical speech and text pipelines with clustering-based workflow automation. At SolarEdge I led fleet data: time series forecasting, survival modelling, anomaly detection, and root cause attribution across millions of devices.

The work has always pulled toward the research end. These days that means building deep learning architectures from scratch (ResNet, ViT, Mask R-CNN, YOLOv8, SimCLR, and the transformer forecasters PatchTST, iTransformer, and TimeMixer) and benchmarking them against published results. Most recently: long-horizon forecasting on the ETTh1 benchmark, lesion segmentation on ISIC 2018 medical imaging data, 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, Time Series & Deep Learning

Time series forecasting with transformers
PyTorch · ETTh1 · Kaggle T4
PatchTST, iTransformer, and TimeMixer built from scratch and evaluated on ETTh1 at four horizons against a DLinear-style linear baseline, with a unit-test suite covering every model. Reconstruction-error anomaly detection reports per-type precision, recall, and F1.
3 architectures from scratch PatchTST MSE 0.398 at H=96 4 horizons
Architecture implementations
PyTorch · torchvision · timm · CIFAR-10 · MNIST
ResNet-18 and Vision Transformer (ViT-Tiny) built from scratch in PyTorch, including multi-head self-attention and transformer encoder blocks, plus a convolutional network implemented from scratch in NumPy. SimCLR contrastive pretraining reproduced from the original paper, confirming its label-efficiency results; augmentation and temperature ablations quantified each component's contribution.
ResNet-18 93.4% ViT-Tiny 86.7% SimCLR 68.2% linear eval 61.6% at 1% labels CNN in NumPy 90.9% MNIST
Medical image segmentation
Mask R-CNN · ISIC 2018 · Kaggle T4
Mask R-CNN for skin lesion segmentation on ISIC 2018 Task 1.
Jaccard 0.782
Object detection
YOLOv8 · PCB defect dataset · Kaggle T4
YOLOv8n fine-tuned on a public PCB defect detection dataset using a pretrained backbone with domain-specific augmentation.
mAP@0.5 0.9896
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 pipelines monitoring millions of inverters and battery units at 24-hour latency; drove anomaly detection and root cause attribution at scale across noisy, correlated multivariate sensor signals.
  • Led failure and survival analysis across device fleets; improved preemptive failure detection by 25% and reduced maintenance costs by 20%.
  • Represented fleet data on a cross-departmental tiger team (20+ people) resolving the most severe inverter malfunctions; drove root cause analysis and A/B tests staged from thousands to hundreds of thousands of live devices, across 3 models with millions of units each.
  • Automated warranty cost projection models, reducing processing time by 90% and direct costs by 15%.
  • Led log and telemetry analysis to identify critical software bugs, achieving 98% detection accuracy.
  • Built a real-time thermal estimation system for inverter devices, cutting overheating incidents by 30%; reduced product failure rates by 60% through systematic root cause analysis.
  • Built a Python/Streamlit SQL interface so non-analyst managers could query fleet data without analyst support.
  • Built multilingual translation pipelines for thousands of recorded customer-support conversations; validated output quality with domain experts.
  • Mentored a cross-functional team of 4-5; established Agile practices and code quality standards.
GE Healthcare – Life Care Solutions Nov 2019 – Jan 2022
Data Operations Engineer
  • Directed the design of a multi-step pipeline for clinical speech and text data across hospital sites, and the automation of a proprietary tagging tool by intent domain, creating a more intuitive and cost-effective process.
  • Managed the continuous learning cycle of ML/DL models using performance diagnostics against a 10,000+ utterance ground truth set, as well as reviews by the tagging team.
  • Led a team of 2–4 developers and data science interns, plus 10+ remote clinical domain experts.
Zeek Mobile – Marketplace for Gift Cards May 2018 – Jan 2019
Junior Data Scientist
  • Implemented feature selection methods in Python and Node.js to address overfitting.
  • Ran SQL-based exploratory analysis and presented findings via Periscope.
Education
B.Sc. Mathematics
The Open University of Israel
Completed 2021

Part-time while employed full-time.

Contact
adimendelowitz@gmail.com LinkedIn GitHub