Expert-led consultancy connecting Explainable AI, EU-funded innovation, and high‑performance data pipelines. Translating ExLiFT, RAISE, ECML/PKDD, and IEEE BigData experience into actionable architectures for regulated, industrial environments.
Commercial offerings directly grounded in the research and projects showcased below.
XAI Architecture · FinTech
Core value: transforming “black box” models into compliant, explainable assets.
Direct link: built on ExLiFT (FNR Bridges) and SHAP-based explanation research.
Deliverables: technical AI audits, XAI integration into existing ML pipelines, and regulatory-grade transparency reports for FinTech stakeholders.
EU Funding · FNR · Luxinnovation
Core value: navigating the full lifecycle of FNR and Luxinnovation projects.
Direct link: proven delivery within RAISE and FNR Bridges initiatives.
Deliverables: proposal drafting (technical work packages), consortium coordination, and hands-on milestone and impact reporting for EU-style grants.
Energy · NLP · Data Engineering
Core value: high-performance data mining for time-series and textual data.
Direct link: grounded in ECML/PKDD energy forecasting and IEEE BigData sentiment analysis projects.
Deliverables: custom Python/Django data infrastructure, real-time sentiment engines, and predictive forecasting models tailored to operational constraints.
Applied research and consulting across Explainable AI, energy systems, and real-time NLP.
FNR Bridges · ExLiFT
Developed for the FNR Bridges (ExLiFT) program to bring transparency into high-stakes financial decision pipelines.
Engineered SHAP-driven explanation pipelines to refine ML classifications, ensuring financial models are auditable and aligned with transparency regulations.
Key achievement: designed a methodology for Refined Classifications using SHAP and prototypical explanations.
Financial NLP
Multi-class text classification for financial-domain document collections.
Benchmarked and deployed pre-trained language models to automate the categorisation of complex financial documents at scale.
Short-Term Forecasting
Short-term electricity consumption forecasting using datasets of varying temporal granularities.
Applied advanced time-series forecasting to optimise energy distribution and anticipate consumption peaks for grid operators.
Hydro-Electric Analytics
Strategic river basin analysis using long-term hydroelectric generation data.
Used time-series data mining to cluster hydroelectric production patterns, supporting long-horizon resource planning and investment decisions.
Real-Time Social Media
Real-time monitoring of Twitter and social channels for brand and market sentiment.
Designed a lexicon-based “MILD” approach prioritising throughput and stability for high-speed sentiment analysis on big data streams.
Robust Embeddings
Word embedding experiments across datasets of different sizes and characteristics.
Evaluated the robustness of NLP embeddings to ensure sentiment models remain stable and accurate from small pilot datasets up to large-scale production corpora.