Job Summary:
o We are seeking a Data Scientist to build, train, and deploy large-scale self-supervised foundation models for time-series, sensor, vision, and text data.
o This role focuses on applications such as anomaly detection, predictive maintenance, forecasting, classification, and multimodal sensor fusion across industrial and scientific domains.
Key Responsibilities:
o Develop self-supervised and semi-supervised models for time-series and multimodal data.
o Design deep learning architectures including Transformers, CNNs, RNNs/LSTMs, and fusion models.
o Build scalable pipelines for large, time-synchronized sensor datasets.
o Fine-tune foundation models using transfer learning, adapters, and few-shot techniques.
o Evaluate models using statistical, time-series, and business-focused metrics.
o Collaborate with cross-functional teams to translate models into production solutions.
Required Skills:
o Strong experience with time-series and sequential data (univariate & multivariate).
o Hands-on expertise with sensor data (e.g., vibration, temperature, audio, images).
o Proficiency in Python (NumPy, Pandas, SciPy); experience with PyTorch and TensorFlow.
o Experience training models at scale (multi-GPU, distributed training).
o Solid foundation in statistics, linear algebra, optimization, and signal processing.
Nice to Have:
o C++/CUDA for performance optimization.
o Experience with multimodal or foundation model architectures.
o Exposure to generative or diffusion models.
Qualifications:
o MS or PhD in Computer Science, Data Science, AI, or a related field.
o 3+ years of hands-on experience in data science or machine learning.