From being handed the spec to owning it

I've loved machine learning since school. I studied engineering anyway, because my dad told me to be different. The road back to ML ran through every job below, and the jobs split cleanly in two: the ones where the work arrived defined, and the ones since ANU, where deciding what to build is the work.

Engineering: learning to think in structures

My degree was in automobile engineering at Manipal, on the mechanical side. It taught me to think structurally: to break a system into parts, understand how each part loads the others, and adapt when one of them gives. That habit has outlasted everything else I learned there.

I had been coding since high school, though, and near the end of the degree I moved towards computer science one step at a time: projects first, then internships, including an ML pipeline for early breast cancer detection, and eventually a job.

Cognizant: proof that the switch was real

My first job, straight out of college, building full-stack .NET and React systems, including a logistics system for hospitals running COVID-19 operations. Its real value was what it proved: that an engineering graduate could make the move into software. I was wanted, I could build, and I did well, all before AI tools existed to help. It is where I earned my first full-stack chops.

Accenture: how real ML gets done

I had loved machine learning since school, so when the chance came to work with data, I jumped. Accenture trained me as a data engineer (Hadoop, Scala, PySpark, Cassandra) and put me inside a real production environment: Agile sprints, a stand-up every morning, stories and story points, and seniors who were genuinely great to learn from.

I built the MLOps pipelines behind Ingrain, an AIOps tool for automated IT ticket resolution, along with its explainability and reporting. That is where I learned why data and pipeline design matter as much as the model, the language of real ML workflows, and how a strong team actually works together.

THEN ANU, AND AI

Up to this point, the work arrived defined. Someone decided what to implement, and my job was to implement it well. I did.

ANU: talking at their level

Alongside the Master's, I am a Student Ambassador for the College of Systems and Society, talking to prospective students at recruitment events and school visits. The lesson there was simple and harder than it sounds: talk at their level. A student should walk away with a sparkle about what they could study, not a frown of worry about whether they are good enough for it.

Eccoi and Haizea: nobody hands you the spec

My two jobs since ANU are new ground in both directions: systems engineering at Eccoi, where I architect sovereign AI, and deep learning at Haizea Analytics, where I map Australia's tree canopy from satellite imagery. Both began after AI changed how software gets built.

The difference is independence. When I ask what to implement, the answer is that it is mine to decide. I am in the driver's seat. I learned agentic frameworks along the way, and the work became human-in-the-loop: the tools do more of the building, while the judgement about what to build, and whether it is right, stays with me.

The thread

Engineering taught me structure. Cognizant proved I could build. Accenture taught me how real machine learning runs. The jobs since ANU are teaching me the part no one can hand over: deciding what should be built at all.