A quiet introduction to the humans behind Golden Ratio.
Llohann Dallagnol Sperança
Founder, Golden Ratio
I'm a mathematician by training — a PhD and a postdoc in differential geometry — and a data scientist by practice, working where research meets delivery: natural-language processing, retrieval, and the quiet engineering that decides whether a model is ever actually used. I teach as a professor at UNIFESP in Brazil, and I supervise master's research projects at Imperial College London — fourteen across the last two cohorts, five of them with distinctions — on retrieval, extraction and NLP over corporate sustainability disclosures. I started Golden Ratio to do two things at once: guide organisations into the AI era, and run a company that is itself the example, with machines doing most of the building. I build AI systems designed to be adopted, understood and maintained by the teams meant to run them, for clients including the Leonardo Centre on Business for Society. Most AI initiatives don't fail because of flawed algorithms; they fail on the day the only person who understands the system is away.
Mojtaba Nouri
Consulting Member
I'm an economist by training and a strategy researcher by practice, doing a doctorate in management in London on how cooperation and innovation shape a firm's sustainability performance, after a research master at Bocconi in Milan. My work sits where strategy meets computation: how much a firm should explore before committing, and how to tell learning from luck. That is game theory and multi-armed bandits on paper, and simulation and machine learning the moment the paper version stops helping — most recently supervising a master's research project on multi-agent reinforcement learning, where agents learn how much to contribute to a shared pool and what their exploration costs them. At Golden Ratio I bring that to problems where the answer has to arrive as something usable, not a model on a whiteboard.
Sanjali Vats
Consulting Data Scientist
I'm a data scientist working in healthcare and pharma analytics, and with Golden Ratio on the part of AI I care most about: retrieval. Getting the right document in front of the right model at the right moment is most of what makes an AI system useful, and most of it is quiet work. I build vector search that returns sensible results, ground chatbots in a client's own documents rather than in the internet's average opinion, and pull text out of the files that resist being read. I like the problems that arrive half-defined, where the question is vague, the corpus is a mess, and the answer has to exist by the end of the fortnight.
Bharat Singla
Consulting Data Engineer
I'm a solution architect: six years of making AI and data platforms work at a size where they stop being demos, most recently leading a team that builds them end to end — the pipelines, the lakehouse and warehouse designs, the feature store, the recommendation and forecasting models on top. What I like is the awkward middle of the job, where a vague requirement turns into a data model nobody has to apologise for. With Golden Ratio I built the systems that read incoming work, judge what is worth our time and draft the first version of a reply. My days have been Bayesian modelling, transformers, and a great deal of SQL, roughly in that order.
