Available for Work

Engineering Excellence through Advanced Simulation, Data Analysis and Process Optimisation

Dr. Kieran Jervis

I’m Dr. Kieran Jervis, founder of ChemDigital Ltd., a digital engineering consultancy delivering high impact solutions across chemical simulation, optimisation, and industrial data science.

With a PhD in Chemical Engineering and a proven track record across industry and consulting, I combine domain expertise with full stack technical delivery. I lead projects from proposal through to deployment; covering solution architecture, CTR planning, stakeholder engagement, technical development, and delivery management.

Whether building simulation engines from legacy spreadsheets, deploying scalable cloud services, or applying machine learning to reduce operational costs, I design solutions that are technically sound, intuitive to use, and scaled to align with business value.

I work independently or as part of larger teams, always focused on clear communication, transparency, and turning complex challenges into elegant, effective tools.

Services

At ChemDigital Ltd., I offer a flexible, client-focused approach, working directly with me ensures clear communication, technical clarity, and a strong understanding of your challenges from day one.

Digital Transformation & Automation

Optimise and modernise workflows by transitioning legacy systems to cloud-hosted solutions. Whether Azure Durable Functions, REST APIs or other preferred stack, I deliver practical, tailored improvements that streamline operations, automate key processes, and create scalable, user-friendly systems aligned with your needs and infrastructure.

Advanced Chemical Process Simulation

Design and develop intuitive, scalable simulation frameworks in Python that provide actionable insights into system dynamics, performance, and efficiency. These custom solutions help optimise your processes, resolve complex challenges, and support strategic decision-making with precision.

Data Analysis & Machine Learning

Transform raw data into actionable insights through advanced analysis and machine learning techniques. With expertise in data exploration, regression, and classification, I can help your business uncover hidden trends, optimise operations, and implement scalable, results-driven solutions.

Process Optimisation & Engineering Design

Achieve energy efficiency, cost reduction, and sustainability goals with tailored engineering solutions. I specialise in multi-objective optimisation, delivering designs and process improvements that align with your unique multivariate operational needs.

Portfolio

I am continuously growing my technical skill set to meet the changing digital demands of industry. My domain expertise and project experience enable me to quickly adapt to new challenges, integrate into active teams, and consistently deliver impactful results. Below are highlights from recent work across engineering and digital innovation. For additional projects, please see my LinkedIn

High-Performance Oil Transient Hydraulics Simulation Platform

Pontem, US

Results: Delivered a production-grade oil transient simulation platform in 4 months, achieving 10–50× faster analysis than commercial software, deployed on Azure for scalable, web-based user enablement.

Problem: Traditional commercial tool workflows (e.g., AFT Impulse) were costly and lacked agility for oil system transients. Operational engineers could not run large-scale emergency shutdown (ESD) analyses independently or on demand, limiting operational decision-making and risk planning.

My Approach:

Identified and adapted a high-performance vectorised Method of Characteristics (PTSNET) open-source codebase, replacing a slower, less scalable proposal
Extended the framework to model oil systems with advanced controls: feeds, flow control valves, check valves, back-pressure terminals, and dynamic friction modelling
Removed MPI dependency to enable cloud-native deployment on Azure
Developed a robust model parser to convert AFT Impulse models for immediate use
Built a user-friendly FastAPI web application with seamless front-end integration
Engineered a parallel fan-out architecture to batch-process multiple simulations concurrently, achieving turnaround times comparable to a single simulation run
Coordinated front-end collaboration to deliver an intuitive user experience

Impact:

Enabled engineers to independently run water hammer analyses at scale, reducing reliance on expensive licenses
Cut simulation turnaround times by 10–50× compared to commercial tools
Positioned the client to model and assess midstream system responses to ESD for both design and operational risk planning
Established a solid foundation to support future SCADA system integration, enabling near real-time scenario modelling for continuous asset health monitoring.

Python • FastAPI • PTSNET • Azure •

JSON/REST APIs • Asynchronous programming

Predictive Maintenance for Milling Equipment (Personal Development Project)

📄View Report

Python • scikit-learn • pandas • data pipelines • feature engineering • Random Forests

ChemDigital Ltd.

Results: Demonstrated a 25× reduction in predicted tool failures and optimised maintenance scheduling through a scalable machine learning pipeline, built as a personal project.

Problem: Diamond-tipped milling machines often experience unpredictable tool-tip failures, leading to costly unplanned downtime and inefficient preventive maintenance. I wanted to explore how machine learning could improve maintenance decisions in a typical industrial environment.

My Approach:

Collected and generated realistic datasets to simulate historical operational data
Engineered robust features representing tool wear and machine load signals
Developed and validated a Random Forest classification model with high recall and precision
Tuned hyperparameters and applied cross-validation to ensure generalisable results
Prototyped a simple scheduling tool to translate predictions into maintenance intervals
Documented the workflow and visualised results for clear communication and reproducibility.

Impact:

Achieved a simulated 25× reduction in tool failure instances compared to a naive preventive maintenance policy
Demonstrated the viability of machine learning for predictive maintenance in industrial settings
Created a repeatable pipeline design that could be adapted to real-world datasets
Demonstrated skills in data-driven engineering decision support through a realistic industrial scenario.

Hydrogen Project pre-FEED Feasibility and Optimisation Platform

Wood Plc, Digitial Consulting, UK

Results: Transformed a static Excel-based feasibility workflow into a scalable, user-friendly simulation platform, significantly reducing study turnaround time and enabling faster, more confident investment decisions for hydrogen and renewable energy projects.

Problem: The existing Excel-based feasibility approach for hydrogen and renewable energy systems was error-prone, difficult to maintain, and lacked scalability. Individual team members maintained separate copies of system models built in excel, creating inconsistencies and version control issues. The disconnected approach could not support multi-scenario analysis or collaborative workflows, which delayed evaluations and hindered stakeholder engagement.

My Approach:

Led the end-to-end transformation of the Excel-based workflow into a modular, Python-driven dynamic simulation platform.
Engineered a flexible architecture where models could be assembled from typical energy–chemical system components (e.g., hydrogen production via electrolysis, ammonia synthesis, battery storage, material-energy-flowlines).
Integrated CAPEX and OPEX calculations to unify economic and technical feasibility analysis.
Built into an in-house modular GUI that allowed users to intuitively construct and edit system models.
Directed interdisciplinary collaboration between back-end engineers, GUI developers, and analysts.
Established a back-end for secure data sharing and consistent model versioning across users.
Created documentation and workflows to support onboarding, reuse, and extension across projects.

Python • scipy • pandas • simulation engine design • ODE systems • System dynamics • JSON/CSV-based data pipelines • REST APIs
Azure deployment • Versioned model storage • Process economics

Impact:

Enabled scalable, multi-scenario feasibility studies for hydrogen, ammonia, battery, and renewable systems
Reduced time-to-insight, helping accelerate internal decision-making and unlock funding
Replaced error-prone spreadsheets with a flexible, maintainable, and modular platform
Empowered users to build and adapt models using domain-familiar building blocks, improving engagement and usability
Streamlined prototyping workflows for hydrogen derivatives projects and de-risked early-stage concept development.

Universal Compartment Modelling Tool for Chemical Engineering

📄 View Thesis

Industrial PhD, The University of Leeds

This research introduced CompArt, a universal compartment modelling tool for chemical engineering unit operations. The framework incorporates phenomenological models of single- and multiphase systems, including mass transfer, reactions, phase transitions, and convective flow.

The project addressed the widespread challenge of bespoke and error-prone model development in the field. CompArt introduced a universal input language, an interpretation algorithm for generating ODE systems, and integration with advanced numerical solvers. By encapsulating complex models within a consistent structure, the tool eliminates the need for custom coding and allows engineers to focus on model development rather than implementation.

The system was validated against 20 benchmark models, demonstrating its ability to simplify and standardise the compartment modelling process for complex chemical systems.

Key Achievements:

Developed a universal framework for building, solving, and validating compartment-based dynamic process models
Streamlined simulation of stiff and non-linear chemical systems, improving modelling efficiency
Enhanced accessibility of advanced modelling techniques by enabling model construction without specialist coding skills