Manav Arora.

What needed to change

For some large tables, preparing metadata for just two could take an entire developer day. Engineers had to repeat detailed configuration work for every table.

What I built

I developed reusable metadata-generation workflows and API templates that produce the configuration needed to onboard source tables.

Why this approach

I made the preparation reusable across tables, so the same workflow could handle repeated setup. The value is the engineering time returned as more tables are added.

What changed

The workflow now generates metadata for 20 tables in two minutes. That timing covers metadata preparation; it is one part of the wider ingestion process.

Microsoft Fabric · Python · Metadata automation

What needed to change

Code alone does not tell a Business Analyst or a new engineer why a notebook exists or how it fits into the work around it. Teams need an understandable starting point.

My contribution

I architected an AI-powered documentation workflow and progressed it through approval and manual testing, alongside the wider Copilot implementation work I lead.

The design focus

The goal is documentation that connects purpose, processing and outputs. It should help people understand a system before they need to work through its implementation.

Where it stands

The solution is approved and manual testing has been successful. Production rollout is next; time saved and ongoing adoption still need to be evaluated.

Microsoft Fabric · Copilot · AI-powered documentation

The challenge

Platform changes and new AI capabilities need clear answers about quality, security, approvals and operational responsibility.

My contribution

I architected a CI/CD approach covering automated quality and security checks, environments and release controls. I also designed an approach to AI-use security and governance.

Why this approach

I designed checks and release controls as part of the delivery path, giving teams a repeatable approach to consider before implementing platform changes.

Where it stands

The CI/CD design is complete, with selected automated checks trialled. The AI security architecture is at design stage. Operational benefits still need to be validated during implementation.

CI/CD · Automated checks · AI governance · Solution design

Where it started

I experienced the uncertainty around finding suitable jobs and understanding how work rights affect opportunities. Other international candidates face the same barriers, while employers need clarity too.

What I’m building

Linzara is a visa-aware jobs platform. As founder, I bring together customer discovery, solution design, the application, cloud infrastructure and AI-assisted features.

Thinking beyond my own experience

I wanted to address uncertainty on both sides of hiring. The product brings skills and work-rights information into the picture before someone applies, helping candidates and employers make more informed choices.

Where it stands

Linzara is a working product with early user testing and an approved university pilot pathway. Hiring outcomes and sustained adoption are still being established.