Question 1

What kind of work do you take on?

Accepted Answer

Research collaborations, founding engineer roles and ambitious projects in ML and systems. Most of my client work is AI products that have to hold up in production: voice, retrieval, model training and the infrastructure beneath them.

Question 2

Where are you based?

Accepted Answer

London, United Kingdom. I work remotely with teams anywhere.

Question 3

How does a project with you run?

Accepted Answer

Directly with you. We scope it and agree an estimate, I build in short cycles with demonstrations along the way, and your feedback steers each round.

Question 4

Can I see your code before we talk?

Accepted Answer

Please do. GrahaOS, SocioMemory and the terrain segmentation work are on GitHub, and the BOOMI models are on Hugging Face.

Question 5

Can I reproduce the SocioMemory numbers?

Accepted Answer

Yes. It ships with memorybench, a TypeScript benchmarking framework released alongside the engine so every published figure can be checked independently.

Question 6

Which language do you reach for first?

Accepted Answer

Rust, most of the time. Sociobot is about 170,000 lines of it. Python for research and model work, and C with assembly when the job is the machine itself, as in GrahaOS.

Question 7

What did you study?

Accepted Answer

A BE in Computer Science and Engineering at University College of Engineering, Osmania University, from 2022 to 2026, graduating with a GPA of 8.7 out of 10.

Question 8

What is the best way to reach you?

Accepted Answer

The form at the end of this page comes straight to me. LinkedIn and GitHub work too.