Ross Murphy

Currently

• I'm building OneContext. We make production grade RAG (retrieval augmented generation) pipelines.
• I'm interested in whether humans and/or machines can reason in natural language, and linguistic relativity in general. Some thoughts on that here.
• I'm also interested in / worried about whether language models can learn (directly, or indirectly) steganographic collusion techniques to successfully evade oversight.

Previously

• I have an MSc in Data Science & Machine Learning from UCL, and an MA (Oxon.) in Economics & Management from Oxford.
• I graduated from UCL on the Dean's List (top 5% of Computer Science Department). While at Oxford, I was awarded the Nind Scholarship.
• While at UCL I did research with The UCL Sustainability and Machine Learning Group and The Mercedes AMG F1 Team. I was very fortunate to be advised by Marc Deisenroth.
• I've spent some time developing and trading systematic strategies in volatility products at a quant hedge fund, and at a large bank.
• I've worked as a consultant machine learning engineer for a couple of interesting ML startups doing smart things with LLMs. For example.

If there's anything else you'd like to know, or if we share similar research interests, reach out!