Software for organizations working with grants
This work is for grant-technology startups, funding platforms, advisory organizations, and teams operating public-funding programmes. It is not a grant-writing service for an individual applicant.
Grant products have to turn fragmented programme information into decisions people can act on. Opportunities are published across different sources, eligibility rules are embedded in long documents, and similar ideas are described with different terminology across sectors and languages. A useful system must reduce that complexity without hiding the evidence or expert judgment behind a recommendation.
What I can help build
Grant matching and ranking
Match company or project information with relevant programmes while controlling noisy results and preserving decisive eligibility or thematic concepts.
Taxonomies and concept models
Design hierarchical categories, synonym maps, weights, and relationships that reflect how experts distinguish funding opportunities.
Programme information extraction
Turn calls, guidance, criteria, and forms into structured information that products and operational workflows can use.
APIs and workflow software
Deliver maintainable services and internal tools that connect matching, review, recommendations, and accountable human decisions.
The hard part is modelling the decision
The visible feature may be a search box, ranking, recommendation, or application workflow. The difficult work happens underneath:
- defining what makes a programme genuinely relevant rather than merely similar in vocabulary;
- distinguishing hard eligibility constraints from softer thematic fit;
- deciding which document sections and concepts should influence a result;
- handling multilingual terminology, synonyms, abbreviations, and sector-specific language;
- choosing category boundaries that are detailed enough to be useful without becoming impossible to maintain;
- showing why a result appeared and where expert review is still required.
Generic semantic similarity or an undifferentiated list of keywords rarely captures all of these constraints. The right design may combine deterministic rules, taxonomies, weighted concepts, natural-language processing, machine learning, and human review. The method should follow the decision—not the other way around.
Research and expert review belong in the delivery
For grant-related products, taxonomy research and evaluation are part of product development. I work with the people who understand the programmes to compare alternative structures, test representative cases, inspect costly mistakes, and determine which distinctions the system must preserve.
That process can lead to a custom API, an internal review tool, document-processing infrastructure, workflow automation, or a combination of these. The objective is not to remove professional judgment. It is to make the surrounding information easier to process, compare, and act on.
Selected work
For a European grant-technology startup, we delivered a multilingual matching API alongside taxonomy research, product recommendations, and iterative evaluation using test data with the client’s domain expert.
Read the anonymized grant-matching case study →
A diagnosis-first engagement
An engagement begins with the actual decision, users, source material, error costs, and review process. From there, I can identify whether the useful intervention is better information architecture, a matching algorithm, document extraction, workflow redesign, or custom software—and build only what the operating model can support.