Senior Data Scientist

Aker BP | 09/2025 – Present

AI@Core / AI Pod – Production Anomaly Detection for Well & Equipment Monitoring

  • Core data scientist responsible for designing, implementing, and validating an end-to-end time-series anomaly detection framework across well production and equipment monitoring, an executive-sponsored AI@Core pilot project evaluating AI capability for real production use cases, with successful pilots progressing to production deployment.

  • Project selected to scale further under the AI Pod campaign targeting large-scale rollout at production maturity, with use cases mapped to key Yggdrasil monitoring anomalies including scale build-up, pressure deviations, productivity tracking, and valve health.

  • Developed physics-informed learned representations encoding thermodynamic, multiphase flow, and hydraulic domain knowledge into interpretable feature spaces from multivariate sensor data, enabling robust detection of fouling, scaling, and equipment degradation across diverse operating conditions.

  • Built and benchmarked anomaly detection approaches including deep learning classifiers, attention-based architectures, and zero-shot foundation models, validated on both public industry benchmarks and real operational data.

  • Designed evaluation frameworks incorporating trend-aware detection scoring and failure-mode analysis, validated against confirmed maintenance events.

Evaluation of NVIDIA NV-Tesseract-AD Foundation Model

  • Led hands-on technical evaluation of NVIDIA’s time-series foundation models for anomaly detection across four model iterations (transformer-based classifier, diffusion-based, zero-shot forecasting, DARR mode), serving as primary technical point of contact in direct collaboration with NVIDIA’s model engineers, facilitated through Cognite. Use case results adopted by Cognite’s Dune team for their application demo at NVIDIA GTC 2025.

  • Systematically benchmarked NVIDIA models across diverse industrial data types (physics-engineered features, raw noisy sensor signals, sparse data, and multivariate well data), demonstrating how physics-informed domain knowledge improves model robustness and reduces configuration sensitivity.

  • Developed novel evaluation methodology including error derivative analysis to handle multiple anomaly dynamics, demonstrating that understanding evaluation metrics is critical for deploying zero-shot models on noisy industrial production signals.

  • Re-implemented NVIDIA’s transformer architecture as a custom multivariate version trained on domain data, achieving comparable performance to their pre-trained model, proving feasibility of in-house alternatives when vendor models don’t cover operational needs.

  • Provided direct technical feedback to NVIDIA on performance findings, identified limitations, and reported bugs, informing anomaly detection scoping for the Yggdrasil production monitoring platform.

AI Foundation Platform Development (Software, Infrastructure, Agents)

  • Core member of the AI Execution & Data Science Team building Aker BP’s enterprise AI foundation for development, deployment, and lifecycle management of AI agents and copilots.

  • Designed and implemented platform architecture covering secure data pipelines, compute management, orchestration, and end-to-end CI/CD for AI systems.

  • Developed reusable ingestion, enrichment, and indexing pipelines using Azure AI Search and Fabric OneLake with strict governance, RBAC, and auditability.

  • Built SDKs, developer tooling, templates, and automated onboarding workflows enabling consistent and repeatable deployment of AI systems across teams.

  • Contributed to production governance standards including model evaluation, Responsible AI compliance, security validation, monitoring, and lifecycle management.

AI-Augmented Engineering & Agentic AI

  • Contributed to defining and shaping best practices for incorporating agentic AI and LLM-based tooling into daily workflows, with a focus on correctness, maintainability, and production suitability.

  • Advised teams on when and how AI-assisted coding can be effectively applied, and where traditional engineering approaches are required, ensuring AI tools are used as accelerators rather than decision-makers.

  • Involved in strategic collaboration between Aker BP and Microsoft on hypervelocity and spec-driven coding initiatives, providing hands-on evaluation, technical feedback, and guidance on practical adoption of AI-assisted development.

  • Evaluated AI-generated code paths, workflows, and design patterns against engineering standards, reliability requirements, and system-level constraints.

  • Applied disciplined judgment in selecting and combining physics-based models, statistical methods, machine learning, and AI tools based on problem structure, risk profile, and production constraints.

Lead R&D Engineer

Marwell Tech AS | 03/2025 – 06/2025

Physics-Based Modeling & Experimental Systems

  • Designed and executed controlled experimental methodologies to characterize dissolution behavior and performance of downhole components under realistic operating conditions.

  • Developed and calibrated physics-based models using experimental data to simulate dissolution dynamics of proprietary alloy systems.

  • Translated experimental observations into simplified, interpretable models suitable for prediction and engineering decision support.

Data Analysis, Validation & Decision Support

  • Led in-house testing campaigns for ExxonMobil’s Guyana project, managing the full workflow from experimental design and data acquisition to analysis and interpretation.

  • Analyzed complex, high-variance experimental datasets to identify key drivers of system behavior and performance limits.

  • Produced technical reports, test summaries, and product documentation to support internal decisions and external stakeholders.

Cross-Functional Engineering Collaboration

  • Worked closely with suppliers, manufacturers, and field engineers to iterate on design choices based on empirical data and operational constraints.

  • Ensured compliance with industry standards and internal quality protocols across experimental execution, data handling, and failure analysis.

Research Fellow (PhD)

University of Stavanger, Norway, Department of Energy and Petroleum Engineering | 08/2021 – 03/2025

Computational Modeling & Multiphase Flow Analysis

  • Conducted research on fluid displacement and multiphase flow dynamics in confined geometries, focusing on nonlinear behavior, instability mechanisms, and regime transitions.

  • Developed and applied computational fluid dynamics (CFD) models to simulate complex multiphase systems governed by coupled nonlinear equations.

  • Performed mathematical analysis and scaling studies using dimensionless parameters to classify flow regimes and identify dominant physical drivers.

Model Validation, Data Analysis & Uncertainty Reasoning

  • Designed experimental methodologies to validate numerical models and assess model fidelity against controlled observations.

  • Analyzed high-variance experimental datasets to extract robust signals, quantify uncertainty, and evaluate model assumptions.

  • Built structured datasets combining experimental and simulated data to support systematic comparison and interpretation.

Research Delivery

  • Integrated experimental findings and numerical modeling to reason about complex system behavior under incomplete and imperfect information.

  • Collaborated with multidisciplinary teams of researchers, engineers, and technicians to deliver project milestones within time and resource constraints.

  • Co-authored multiple peer-reviewed publications in high-impact journals.

Research Fellow

Polytechnique Montréal, Canada, Department of Mechanical Engineering | 01/2018 – 12/2020

  • Led a research project on autothermal reforming (ATR) for synthesis gas production, focusing on combustion processes, efficiency optimization, and emissions reduction.

  • Designed, built, and operated pilot- and bench-scale experimental systems for ATR using CAD-assisted design, supporting feasibility assessment and scale-up considerations.

  • Developed and applied advanced experimental diagnostics, including optical and thermal measurement techniques, to characterize multi-gas combustion behavior and system performance.

  • Translated experimental observations into quantitative models and engineering insights to support decision-making and downstream application.

  • Collaborated with international partners, co-authored peer-reviewed publications, and presented research findings at international conferences.

Graduate Research Fellow

Georgia Institute of Technology, USA, Department of Aerospace Engineering | 05/2016 – 12/2017

  • Conducted numerical and computational studies of plasma-assisted combustion using CFD, DNS, and multi-physics simulation to analyze ignition behavior, kinetics, and efficiency under extreme conditions.

  • Developed high-performance scientific computing tools using Fortran, MATLAB, and parallel computing to simulate multi-scale systems.

  • Co-authored peer-reviewed publications and presented results at international conferences.

  • Parallel research in early deep learning for aerodynamic prediction, developing CNN-based surrogate models for airfoil lift estimation.

  • Designed data pipelines, feature extraction, and model training workflows for high-dimensional aerodynamic datasets, producing highly cited results.

Graduate Teaching Assistant

Georgia Institute of Technology, USA, Department of Aerospace Engineering | 08/2015 – 05/2016

  • Led laboratory sessions and lectures in control systems, supported by numerical simulation and experimental design.

  • Refined and implemented lab materials demonstrating key principles of control theory, numerical simulation, and stability augmentation systems for aircraft and spacecraft.

  • Assessed student performance through grading of homework, exams, and lab reports, providing detailed feedback to support learning and skill development.

Undergraduate Research Assistant

Georgia Institute of Technology, USA, Department of Aerospace Engineering (Rotational Program) | 01/2013 – 05/2015

  • Participated in a multi-year rotational research program across combustion, high-power electric propulsion, and aeroelasticity, supporting experimental design, data analysis, and modeling.

  • Contributed to experimental and analytical studies presented at international conferences.

DAAD Research Fellow

TU Berlin, Germany, Institute of Fluid Dynamics and Technical Acoustics (ISTA) | 05/2014 – 08/2014

  • Conducted experimental research on pulse detonation combustion systems, applying optical diagnostics and collaborating in an international research environment.