Data Scientist, Engineer

This is a fully remote position, open to applicants in Ukraine.

📋 Description

• Evaluate high-dimensional sensor and feature datasets utilizing techniques such as UMAP, t-SNE, PCA, and others.

• Detect clusters, anomalies, blind spots, distribution gaps, and mismatches between classes or environments.

• Assess dataset shifts, domain drifts, sparsity issues, and representation collapse.

• Conduct data analysis in accordance with classical machine learning models, including XGBoost, SVR, k-NN, and tree-based models.

• Provide analytical support for deep learning models like CNNs and Transformers.

• Examine embeddings, confusion matrices, and model failure patterns to trace errors to data-related issues.

• Investigate imbalanced data, noisy sensor signals, mislabeled samples, and ambiguous cases.

• Create strategies to enhance weakly labeled or unlabeled data using clustering and pseudo-labeling methods.

• Execute data mining on extensive collections of field data to uncover insights and patterns.

• Develop methodologies for transforming noisy or partially verified data into high-quality validated datasets.

• Convert exploratory findings into actionable recommendations for data filtering, relabeling, or new data collection.

• Promote and implement data-centric enhancements to improve model robustness.

• Collaborate closely with engineering teams to integrate improved data workflows into machine learning pipelines.

⛳️ Requirements

• Master's or PhD degree in Data Science, Computer Science, Applied Mathematics, Statistics, Physics, or a related discipline.

• Over 2 years of practical experience working with machine learning datasets.

• Proficient in handling time-series, sensor, image, or video data.

• Strong programming skills in Python and experience with libraries such as NumPy, pandas, matplotlib, and seaborn.

• Familiarity with dimensionality reduction and representation analysis techniques like UMAP, t-SNE, and PCA.

• Comprehensive understanding of machine learning principles, model evaluation, and diagnostics.

• Experience in supporting both traditional machine learning and deep learning initiatives.

• Nice to Have: Background in working with sensor data including radar, magnetic, environmental, 3D, or IoT datasets.

• Knowledge of scikit-learn preprocessing workflows.

• Experience managing imbalanced datasets, noisy labels, sensor noise, and data drift.

• Understanding of model interpretability, feature importance, and embedding analysis.

• Previous experience collaborating with data annotation and labeling teams.

• Familiarity with tools such as MLflow, Weights & Biases, DVC, or similar platforms.

🏝️ Benefits

• Health insurance from the first day of employment.

• Christmas holidays from December 25 to December 31.

• Collaboration with Superhumans center and Veteran HUB.