← Previous part · Knowledge Base · Next part →

Which of my projects best demonstrates my suitability for this role?

Author: Mustafa Siddiqui · Part: 2 of 3 · Recruiter question: Which of Mustafa's projects best demonstrates his suitability for this role?

Evidence rule: This is my first-person authored source material. SudoChat may summarise it in third person but must not strengthen, exaggerate or invent claims. Limitations and Do not claim sections are hard factual boundaries.

## Why SudoChat alone is not the entire answer
SudoChat is a proof of concept.

If it were my only evidence, it would demonstrate technical curiosity but would not necessarily prove that I could deliver inside a large government organisation.

That is why my other work matters.

### DXC document approval automation proves delivery

At DXC, I volunteered to solve a manual approval workflow that required approximately four hours of repetitive effort.

I investigated the process, designed a SharePoint and Power Automate solution, created a dashboard and wrote user documentation.

The resulting process required approximately eleven minutes of monitoring and oversight.

This proves something that SudoChat alone cannot:

**I can identify a real organisational problem and deliver a solution that produces a measurable improvement.**

It also demonstrates judgement.

I did not force AI into the process.

Conventional workflow automation was the appropriate solution, so that was what I used.

### Government enterprise experience proves I understand constraints

Through Unisys, Nova Systems, Xaana.AI and DXC, I have worked in or supported environments associated with organisations including:

* Department of Home Affairs
* ASD
* DFAT
* Department of Defence
* Services Australia
* Australian Space Agency
* Canberra Hospital

These environments taught me that enterprise engineering involves more than making the technology work.

It also involves:

* identity
* security
* permissions
* existing infrastructure
* approvals
* operational support
* user requirements
* governance
* change management

That experience is important because a Federal Court AI solution would exist inside exactly that kind of wider organisational system.

### Privacy preserving chatbot research proves responsible AI is not just application wording

My work with a behavioural scientist exploring privacy preserving government chatbot architecture provides another important part of the evidence.

The work considered what happens when citizens enter sensitive information such as:

* Medicare numbers
* Centrelink CRNs
* addresses
* banking information
* personal information

into an AI interface.

This encouraged me to think about privacy as part of the architecture rather than simply displaying a privacy warning.

That experience complements SudoChat because it shows that my responsible AI thinking extends beyond prompt design.

### Xaana.AI proves professional AI engineering

At Xaana.AI, I worked on AI document processing and OCR pipelines using technologies including PaddleOCR and OpenCV.

My work reportedly reduced processing time and significantly reduced licensing costs.

This provides evidence that my AI experience is not entirely theoretical or personal experimentation.

I have also worked professionally on AI systems that needed to support real workflows.

## Why this combination matters
If one project must be selected, SudoChat is the strongest match.

But my suitability is better represented as:

**SudoChat**

proves my understanding of modern generative AI and responsible AI design.

**Xaana.AI**

proves professional AI engineering experience.

**DXC automation**

proves practical organisational delivery.

**Government enterprise work**

proves I understand secured and governed environments.

**Privacy preserving chatbot research**

proves that responsible AI is an area I have investigated deliberately.

Together, these create a much stronger case than any single project could.

## Relevance to the Federal Courts
The Court AI Technologist role sits at the intersection of several areas:

**AI**

**enterprise technology**

**users**

**governance**

**security**

**responsible experimentation**

**technical delivery**

SudoChat touches almost all of those areas directly.

Consider a potential Federal Court knowledge assistant.

I would need to think about:

* what information the assistant can access
* which sources are authoritative
* how retrieval works
* how hallucinations are reduced
* how sources are presented
* what happens when evidence is missing
* what users are authorised to access
* how privacy is protected
* what the AI should refuse
* where humans remain accountable
* how the system is evaluated

Those are precisely the questions SudoChat is allowing me to experiment with.

The scale and consequence would obviously be much greater within the Federal Courts, but the engineering principles are transferable.

## What SudoChat reveals about me
The most important thing about SudoChat may not be the technology itself.

I could have produced a visually impressive chatbot that answered every question confidently and called it complete.

Instead, much of my effort has gone into defining:

* evidence
* limitations
* gaps
* source boundaries
* hallucination behaviour
* responsible responses
* what the system must not claim

That says something about the type of AI Technologist I want to be.

I am interested in making AI more capable.

But I am equally interested in defining the boundary between:

**what the system can technically do**

and

**what the system should responsibly do.**

For a Court AI Technologist, I believe that distinction is fundamental.

## Limitations or gaps
SudoChat is a proof of concept.

It should not be represented as a production enterprise AI system.

It does not use Federal Court internal data.

It does not represent the Federal Courts' existing AI architecture.

Its current knowledge base is deliberately curated and relatively small compared with the information environment that would exist in a real Court system.

My direct Copilot Studio experience is also relatively recent.

The value of SudoChat is therefore not that it proves I have already built a production Court AI platform.

It proves that I understand many of the technical and governance questions I would need to address if given the opportunity to work on one.

← Previous part · Knowledge Base · Next part →