Time Series Modeling for Industrial IoT
Robust time series models that move beyond black-box forecasting. Temporally adaptive representations, uncertainty-aware learning, and real-world industrial signal processing.
Research & Capabilities
The frontiers of AI
Our applied research lab bridges the gap between academic breakthroughs and production systems — turning emerging techniques into reliable, deployable advantage.
Research Focus
Reinforcement Learning for Critical Infrastructure
AI-driven control for large-scale power grids within the L2RPN research ecosystem. Structure-aware learning that exploits grid topology to make RL scalable and stable.
Signal Processing and Sensor Analytics
Advanced signal processing for noisy, high-frequency sensor data. Spectral analysis, filtering, feature extraction, and anomaly detection for industrial and embedded systems.
Causal Inference and Probabilistic Modeling
Moving beyond correlation to causation. Bayesian methods, causal discovery, counterfactual reasoning, and intervention analysis for robust decision-making under uncertainty.
Applied Capability
Statistical and Predictive Modeling
Classical machine learning done right. Regression, classification, clustering, and ensemble methods with rigorous validation, interpretability and production-grade reliability.
Marketing Mix Modeling
Quantifying marketing effectiveness across channels. Attribution modeling, budget optimisation, and incrementality testing to measure true ROI and guide spend allocation.
Domain-Specific LLMs
Training and fine-tuning local language models for industry verticals. Optimised for on-premise deployment, data privacy and domain accuracy without external APIs.
Agentic Systems
Multi-step task execution with planning, tool use and self-correction. Focus on reliability and controllability for enterprise deployment.