High-quality computer vision models are trained on high-quality data, which can be hard to obtain. Bifrost uses 3D graphics to generate perfectly-labelled and photorealistic synthetic data.

We have a diverse tech stack, including webapps, 3D applications (Unreal, Blender, Houdini), video streaming servers, ML pipelines. I own the ops and infrastructure challenges, such as delivering our product to air-gapped environments, running GPU-enabled Windows containers, and more.

At some point, Motional became big enough to warrant the creation of an SRE team, and I happily made the switch. I created a central monitoring stack, standardised observability practices across all engineering teams, and babysat several Kubernetes clusters and hundreds of AWS accounts.

Autonomous vehicles encounter unexpected events all the time, so the Fleet team created a Command Center to keep an eye on them and troubleshoot them when they misbehave. I was their first backend engineer, and grew it from a simple monitoring tool into a full fleet management platform.

It later became a key asset for our commercial ventures: I worked with engineers from Uber, Lyft, and Via to integrate our robotaxis into their networks.

Arup started an internal digital transformation movement, and employees with cool ideas were given budget and time to work on them. I helped many teams from the Consulting group turn their ideas into reality.

My favourite project involved modelling the ExperienceLab in VR to convince the MD that it would be worth the investment. Spoiler: it was.

I was president of the student government and was TA for mathematics, algorithms, and biology.

I build tools to improve orchestra management workflows.

I supported administrative and concert operations.

I recruited a new management committee and led several projects.

I recruited members for the board.

I assembled the founding committee and led the pilot fundraiser, which raised $100,000.