Work / Evidence behind the advice

Technical work
behind the advice.

At Camunda, working prototypes helped shape an early AI product direction. At Semaverse, I connected research workflows with source access and permissions. Building a data business also taught me the limits of technical success.

How focused technical investigation supports a decision

01 / Camunda

Prototypes that informed
Camunda’s AI direction.

Consulting through Bold Data · Early AI product and technology work

The product and advisory remit

The product question

Help Camunda explore early AI product capabilities through working prototypes that its product and engineering teams could evaluate.

Technical evidence

I built a documentation and code assistant that accounted for different software versions. Another prototype helped users write and correct expressions in Camunda’s expression language.

For the expression-language assistant, I used a larger language model to generate thousands of training and test examples. I then fine-tuned a smaller model for the task.

How it informed the direction

The prototypes informed Camunda’s early product and technology strategy. They were used internally and subsequently demonstrated at Camunda conferences, giving teams working examples to evaluate and develop further.

02 / Semaverse.ai

Bringing research into
the working document.

Co-founder · Platform architecture and engineering

The founder and partnership remit

The product context

The AI research platform connected source material, research and the documents used for M&A and market intelligence. Product choices included how people worked with the research and which sources they could access.

Workflow and source access

I built a two-way PowerPoint workflow that turned corporate templates into research pipelines and put the results back into branded decks. The working document became part of the research process.

Tenant isolation and fine-grained access controls were part of the platform architecture. Connecting commercial sources involved permissions and authentication alongside retrieval.

Technical contribution

I designed and built research agents, persistent memory, knowledge graphs and cross-source search. Web crawling and document ingestion brought source material into the platform, alongside integrations with commercial data providers.

External data relationships

Semaverse became a PitchBook Premium Connector partner and a Crunchbase MCP launch partner. My integration work also included Tracxn and support for Crunchbase’s MCP and OAuth implementation.

03 / Early Bold Data

The commercial case
for a data platform.

Independent data platform

The business and technical context

The aim was to collect public Amazon product and bestseller data across markets, with operating costs a solo business could support.

Technical evidence

After Holland & Barrett, I independently built a scale-out crawler and scraper running across Google Cloud and AWS. It collected approximately 250 million product extractions across the US, UK, Germany and other markets. This is a count of extractions, not unique products.

Commercial outcome

The platform collected data at scale with low operating costs. I did not turn it into a commercially viable business. Demand and a credible route to revenue belong in the investment case alongside operating cost and technical capability.

Published data and analysis

The public releases cover part of the collection. They are historical datasets; check their collection dates and licence before use.

Explore the published dataset

Read the bestseller analysis

Independent technical practice

Project snapshot · 20 October 2026. Tax Shrink and Llamar are public, BrowseBestsellers is in alpha, and Turmberg remains in playtesting and curation.

Tax Shrink

Public

Tax Shrink, a UK salary and dividend calculator for company directors.

Llamar

Public · Chrome extension

Llamar, a Chrome extension for source-linked web summaries.

BrowseBestsellers

Alpha

BrowseBestsellers, an independent Amazon Best Seller category finder.

Turmberg

Playtesting & curation

A strategy game exploring how generated systems and content work with human design and judgement.

Leadership behind the work

Responsibility for
the wider decision.

My strategic responsibilities have included building internal capability, shaping data strategy and leading an international consulting practice.

Explore my experience

For lessons from these investigations, read the insights.

Start with the decision

What is at stake
for your business?

Considering a technology investment, a change in product direction or a different way of organising your teams? Tell me the decision, what is at stake and your timing.

Discuss the challenge