Time Series Researcher (AI)

Location: Hybrid in Houston or Bengaluru

Company

hirify.global develops Orbital, a physics-informed foundation model for energy operations across oil and gas, refineries, and petrochemicals.

What you will do

• Design and implement foundational time-series architectures for forecasting, classification, anomaly detection, and optimisation- and control-adjacent tasks.
• Develop hybrid models combining classical statistical methods, deep learning architectures, and physics-based constraints such as conservation laws and differential-equation priors.
• Build uncertainty-aware and robust models that handle sensor drift, sensor failure, regime changes, and sparse or delayed ground truth.
• Translate research into production through real-time inference, model deployment, continuous retraining, and integration with downstream agents and optimisation layers.
• Containerise and deploy models with Docker on AWS and Azure, including EKS, ECS, and SageMaker, and build CI/CD workflows for training, evaluation, rollout, rollback, and automated retraining.
• Define back-testing, benchmarking, and validation protocols across datasets, operating regimes, and failure modes.

Requirements

• PhD in Computer Science, Statistics, Applied Mathematics, Physics, or a related field.
• First-author publications in time-series modelling, forecasting, signal processing, or physics-informed machine learning.
• 3+ years of hands-on research experience in time-series or sequence modelling.
• Demonstrated experience with deep learning and probabilistic modelling.
• Expert Python skills with production-grade PyTorch code.
• Experience deploying machine-learning models into real systems.

Culture & Benefits

• Research is evaluated by production impact rather than publication count.
• Principled models, honest benchmarking, responsible shipping, and aggressive iteration are valued.
• Physics, statistics, and machine learning are treated as complementary disciplines.
• The role owns core intellectual property and is responsible for models used in production.