Work / Interactive Professional Portfolio

Interactive Professional Portfolio

A fragmented professional history rebuilt as a governed evidence corpus, then made explorable through Ask Niche Clever.

How the work ran

A sequence of work, not a fixed method.

01

A professional history is not a clean dataset

The project began as a portfolio problem and became a representation problem.

02

From evidence estate to governed corpus

Different sources, dates, claims and maturity states needed to remain distinguishable while becoming usable together.

03

Keeping chronology and provenance usable

The corpus had to support useful answers without turning uncertainty, repetition or later interpretation into false certainty.

04

Learning that cleaner can be worse

A successor became neater but lost answerability, so the previous corpus became a regression benchmark rather than something to discard.

05

Making the corpus explorable

Ask Niche Clever turned the governed record into a question-led public route without making the chat interface the source of truth.

06

Separating the reusable plugin

AskFolio emerged as reusable WordPress software around the interaction problem, distinct from the corpus itself.

How do you turn years of fragmented professional evidence into something another person can actually interrogate and trust?

I began the Interactive Professional Portfolio because a conventional CV or project gallery could only show a small part of the work.

The difficulty was not simply that there was too much material. Professional history does not arrive as a clean dataset. It is distributed across project files, websites, repositories, publishing records, emails, handovers, public posts, current recollection and later summaries. Those sources do not all have the same authority, date precision, authorship or relationship to what was actually deployed.

The project therefore became less about assembling a bigger portfolio and more about building a professional record that could preserve those differences while still being useful.

Current position

V10.1 is the current promoted Interactive Professional Portfolio corpus as of 1 October 2026. The corpus is operational and actively maintained. AskFolio (formerly WP Interactive Portfolio) is a separate reusable WordPress plugin that provides the controlled question interface; it does not create or manage the corpus.

01 / A professional history is not a clean dataset

A normal portfolio usually starts by selecting projects and writing a retrospective account of them.

That works until the record becomes large enough, old enough or interconnected enough that the retrospective account starts hiding important distinctions.

A repository may show that code existed without proving that it was deployed. A role date can establish where I worked without proving what I was doing on a particular day. A later summary may identify a real pattern without proving that I understood the pattern in those terms at the time. A sentence produced during later synthesis should not quietly become a quotation from me.

The Portfolio needed to hold those differences rather than flatten them into one polished career story.

02 / From evidence estate to governed corpus

The underlying evidence came from many different kinds of source, and useful questions cut across them.

The corpus therefore had to preserve things that a conventional professional biography can usually ignore: chronology, provenance, uncertainty, authorship, project maturity, changing interpretations and the difference between a source and what somebody later concludes from it.

It also had to remain selective. The existence of private or sensitive material does not make that material appropriate for a professional corpus or a public interface.

The goal was not maximal disclosure. It was a record strong enough to support proportionate claims and useful exploration without losing the boundary between evidence, recollection and interpretation.

The detailed process used to build and maintain the corpus is part of Niche Clever’s internal working practice. This case study focuses on the problems it had to solve, the distinctions the resulting corpus needed to preserve, and the public systems built on top of it.

03 / Keeping chronology and provenance usable

As the corpus developed, several apparently simple ideas turned out to be separate problems.

Having evidence was not the same as being able to retrieve it at the right level of detail. Saying that an archive or collection existed was not the same as preserving the actual items needed to answer questions about it. Repeating an interpretation across several later summaries did not make that interpretation several times more evidentially important.

Chronology also needed its own discipline. Some questions can be answered from role dates. Others need project state, publication dates, contemporary artefacts or bounded reconstruction. The system needed to be useful at both levels without pretending that a broad date range proves an exact event.

The same applied to authorship and maturity. Automated organisational output can be evidence of a system I operated without proving that I wrote every sentence. Code can be implemented without being deployed. A prototype, an internal operational tool and a live product are different states.

Those distinctions became part of what the corpus had to preserve, not footnotes added after the answer was written.

04 / Learning that cleaner can be worse

One of the most useful lessons came from a version that looked like an improvement.

During the V10 work, an early successor became cleaner and more explicit about governance, but testing showed that it had become worse at answering some concrete questions. It could describe the existence of evidence while no longer preserving enough of the answer-bearing material itself.

Rather than assuming that the newer structure must be better, the previous corpus was turned into a regression benchmark. A successor had to show that it could still recover useful evidence and answer questions at comparable or better resolution.

That changed the definition of progress.

The test was no longer simply whether the corpus was tidier, more elegant or more internally consistent. It was whether somebody could still ask a real question and recover the right evidence, with the right distinctions, at the right level of detail.

05 / Making the corpus explorable

Once the corpus became useful, another problem appeared: nobody should have to read the whole thing in order to benefit from it.

Ask Niche Clever provides a question-led route into the professional record. Visitors can ask about projects, periods, current work, career history or recurring capabilities and receive an answer grounded in the approved Portfolio rather than a generic chatbot biography.

The conversation layer is deliberately downstream of the corpus. Improving the interface cannot repair bad evidence, and a fluent answer does not become authoritative merely because it sounds convincing.

That separation also makes the authored website more useful. The public pages can stay selective and readable, while the larger record remains available for deeper exploration.

06 / Separating the reusable plugin

The interaction problem eventually became reusable software in its own right.

AskFolio is the WordPress plugin that provides the controlled question interface used by Ask Niche Clever. It handles the WordPress-side experience around an approved File Search knowledge base: configuration, public presentation, access states, limits, evidence routes and operational safeguards.

It is deliberately separate from corpus construction.

A site owner can use the plugin with an approved knowledge base of their own. The plugin does not decide what their evidence should contain, build the corpus for them or turn ungoverned material into trustworthy professional knowledge.

That distinction matters enough to keep the product on its own page.

07 / What I did

My contribution spans the evidence, editorial, product and implementation layers.

It has included:

  • reconstructing and reconciling professional evidence across projects and periods;
  • designing the information and corpus architecture needed to make that record usable;
  • defining and reviewing chronology, provenance, authorship and maturity distinctions;
  • testing whether successive corpus versions actually preserved answerability;
  • identifying retrieval and composition failures rather than automatically treating every bad answer as a source defect;
  • directing bounded research where a concrete gap or contradiction needed stronger evidence;
  • shaping Ask Niche Clever as a public exploration route;
  • setting the product direction for AskFolio and directing its agent-assisted implementation and testing;
  • maintaining the boundary between the professional corpus, private source material and the public website.

The important point is that the corpus is not simply content prepared for an AI interface.

It is an information system in its own right.

Where it stands

Interactive Professional Portfolio corpus

V10.1 is the current promoted corpus. It is operational, versioned and actively maintained.

Ask Niche Clever

The Niche Clever-facing conversational interface over the approved Portfolio corpus.

AskFolio

A separate reusable WordPress plugin providing the controlled interaction layer. It remains an active product-development stream.

Corpus-building practice

An internally developed part of Niche Clever’s working practice. Its principles and outcomes can be discussed publicly; the detailed production process remains internal.

08 / What the project demonstrates

The Interactive Professional Portfolio demonstrates that a large evidence estate does not become useful simply because it has been collected or placed behind a language model.

Useful professional knowledge has to retain the distinctions that make it trustworthy: what happened, when it happened, what supports it, what remains uncertain, who authored the language, what was actually deployed, and which interpretations were developed later.

It also demonstrates a broader capability: turning fragmented evidence into a governed knowledge corpus from which trustworthy interfaces, publications and decisions can be produced.

That is the part of the work that extends beyond the Portfolio itself.