Bringing together minds to make the intractable trivial.

Who we are

Our backgrounds are in theoretical physics, quantitative finance and systems engineering. We have built pricing and risk models in investment banks and hedge funds, run data platforms that process tens of gigabytes a day, and shipped production software from mobile clients to network infrastructure. The common thread is a habit of asking what a measurement can actually support before building on it.

How we engage

• Research engagements. A defined question, a dataset or a procurement decision, and a written answer with the evidence behind it. Typically four to twelve weeks.
• Build engagements. A model and the pipeline that runs it, delivered with an evaluation harness, infrastructure code and a handover. Typically three to nine months.
• Advisory. Standing access to the team for organisations that have their own engineers and need a second opinion on method, data purchasing or validation.

What we will not do

We do not clean up data to make it look better than it is, and we do not report a single accuracy number when the distribution tells a different story. If a dataset cannot answer the question, we say so before you buy it.

Interested in working with us?

We are always interested in hearing from researchers and engineers who like problems at the boundary of statistics and software. Write to info@advcipher.com.