Custom algorithms and solvers
Development and adaptation of numerical methods, domain-specific algorithms and solver workflows for engineering problems that require greater flexibility than standard software can provide.
Building Avkalan Labs
Creating a technical organisation around engineering simulation, custom solvers, scientific machine learning and research-led products.
About the project
Avkalan Labs started with the vision of making advanced engineering R&D capabilities more accessible to industry. The aim was to build a technical organisation that could work on complex problems requiring specialised simulation, numerical methods, scientific computing and research-led software development.
The work has grown through collaborations and technical engagements across Asia, Europe and the United States. These projects have involved different engineering domains, but they share a common need: translating difficult physical problems into reliable computational methods and usable engineering workflows.
The broader direction is to contribute to R&D advancement in key technology areas such as semiconductors, aerospace, defence, electromagnetics, thermal systems, advanced materials and healthcare products. My work has included shaping the technical direction of the company, contributing to solver and algorithm development, structuring project execution, building partnerships and identifying how project-specific work can evolve into reusable capabilities.
How it works
The work usually begins with an engineering problem that cannot be addressed adequately through standard software, generic models, or readily available datasets. The first step is to understand the physical system, operating conditions, material behaviour, boundary conditions, available evidence and the decision the engineering team ultimately needs to make.
Depending on the problem, the technical solution may involve developing or adapting a numerical solver, creating specialised algorithms, automating simulation campaigns, generating structured engineering datasets, or building scientific machine-learning models. The work also includes preprocessing, meshing, high-performance and parallel computation, post-processing, visualisation and validation against analytical, numerical, or experimental references.
The final objective is not limited to producing an isolated simulation result. The aim is to translate the technical work into a repeatable capability, such as an internal engineering tool, an automated computational workflow, a validated scientific-ML model, a domain-specific algorithm, or a research-led product that can support future engineering decisions.
Engineering problem → numerical methods → simulation and data → validation → technical capability
What the work produced
The work has established a foundation for developing specialised engineering software and research-led computational systems across multiple technical domains.
Simulation and scientific-ML workflows
Integrated workflows connecting automated simulations, structured datasets, scientific machine learning, surrogate modelling, optimisation and validation against physics-based results.
Cross-domain R&D capability
A technical practice spanning semiconductors, aerospace, defence, electromagnetics, thermal systems, advanced materials and connected products, supported by collaborations across multiple regions.
Let’s talk about what you’re building.
I’m always open to conversations about engineering, computation, deeptech products and mentoring people building or learning in these areas.