Ventures
ChAI
ChAI helps businesses predict and protect against rising raw material costs. It was born out of a simple idea: the tools used by traders in investment banks and hedge funds to anticipate and manage price movements shouldn't be limited to financial markets—they should be available to the businesses operating in the real, physical world.
The price of energy, food, and materials underpins the entire global economy, yet most companies have almost no control over it. They absorb shocks after they happen, or rely on tools that are complex, inaccessible, and built for specialists. ChAI changes that.
Using artificial intelligence, we forecast how these prices are likely to move—turning a deeply uncertain problem into something businesses can see coming and act on in advance. We then provide insurance that pays out when costs rise, allowing companies to protect themselves without needing to navigate financial markets.
For the first time, this makes one of the largest hidden risks in the economy visible, manageable, and directly controllable.
Internews
Internews works to ensure people everywhere have access to trusted, independent information. In an increasingly uncertain world, the flow of reliable information is not a luxury—it is essential to safety, stability, and functioning societies. When information breaks down, so does trust, coordination, and the ability to respond to crises. Internews addresses this at its core.
Working across 40+ countries, it supports local media, civic actors, and information networks to ensure that accurate, relevant information reaches the people who need it most—particularly in fragile and rapidly changing environments.
The result is stronger, more resilient communities that are better able to navigate uncertainty, make informed decisions, and remain connected to the world around them.
Thought Machine
Modern economies depend on banking systems that are reliable, responsive, and able to adapt in real time—yet much of the world still runs on infrastructure built decades ago. Thought Machine replaces these legacy systems with cloud-native platforms, enabling banks to operate with the speed, flexibility, and resilience of modern software.
Its technology now underpins core banking systems at several tier-1 institutions globally.
Informed Actions [inactive]
Emergency response systems depend on making the right decision in seconds—yet are often constrained by limited data and rigid processes. Informed Actions applied machine learning to optimise ambulance dispatch for South Africa’s largest emergency medical service, materially reducing response times.
A clear demonstration of how AI can be embedded into real-world systems to improve outcomes where it matters most.
Invinio [inactive]
Fine wine investment has traditionally been opaque, fragmented, and driven by intuition rather than data. Invinio brought quantitative rigour to the space, combining pricing data across 10,000+ wines with portfolio construction and optimisation tools.
A demonstration of how data and machine learning can bring structure and discipline to previously unstructured asset classes.
Water Purification in the Developing World
Access to clean water is one of the most basic yet persistent global challenges, with direct consequences for health, productivity, and quality of life. This project explored a hand-powered purification device designed for use in environments with limited infrastructure—aiming to deliver safe water where it is needed most.
It reflects a broader idea: that thoughtful engineering and innovation, applied in the right way, can fundamentally transform everyday life for millions of people.
Horrorsight
Most recommendation systems optimise for popularity, not taste—leaving much of the most interesting content hidden. Horrorsight uses AI to uncover cult and underseen horror films, surfacing highly specific styles, atmospheres, and niches.
A demonstration of how intelligent systems can reshape discovery in creative domains.
Many of these ventures draw on machine learning techniques covered in Tristan's free tutorials — including Support Vector Machines and Kalman Filters. See also his academic publications on AI for finance and healthcare.