About

Recently, I started working as a Forward Deployed ML Engineer at Applied Computing. We build and deploy AI models and a reasoning layer that help refineries and plants run better.

Plants and refineries were completely new to me. My PhD and early research were in energy too, but that was seismic data and imaging: a lot of science and not a single valve.

Applied Computing is a startup. From the day I joined, it has been a rollercoaster ride, exciting and a great learning experience. I learned on the job how ammonia and olefins plants work, scoped out the work, understood what the client expected, did research, deployed code, kept it running and wrote it all up.

Six months in, nothing has changed. I still do a bit of everything: client calls, keeping the scope in check, taking research code to production and keeping it running, plus the evals and monitoring around it.

Before this, I built image generation and enhancement pipelines at Jiffy, and before that at Flixstock, with some contract work for other companies on the side.

Lately, I have been curious about owning the intelligence: running open models on a proper serving stack, not just on a single GPU. That means we can deploy them in production for clients who do not want their data going to a model provider, and maybe even use them in our own coding harness. I am trying a few things at work and will write about it soon.

Outside work, I tinker with things and write here when I find the time, which is less often than I would like.

Professional experience

Contract work

Contact

• Email
• LinkedIn
• GitHub
• X