A Look Inside: AI for Investment Insights

Invest Insight
9 followers
I'm Jeff Wecker and I run engineering for Two Sigma. AI has been a core part of Two Sigma's
platform since our founding over 20 years ago. Talent drives our innovative process,
so we hire experts across pure math, statistics, computer science, and of course machine learning
and artificial intelligence. We've always been on the forefront of adopting new technologies. It's
one of the foundational tenets of our business. We really focus on the problem we're trying to
solve and we ensure that we select the right tool for the problem at hand. We can make sure
that we stay ahead of emerging technologies and apply them effectively. AI is only as good as the
data that you feed it. To deliver the more than 300 petabytes of data across the public and private
clouds that our researchers use, we have to invest pretty heavily in developing our data pipelines.
Our business demands perpetual curiosity, data science, and exploration with the latest tools
to ensure we're delivering the best outcomes for our clients. We not only use AI to deliver
predictive insights for our investment management business, we use it all over the company. So
whether that is incident management or code generation or computational research, in terms of
generative AI, public LLMs are remarkably flexible and adaptive tools, but you have to have a clear
understanding of the problem you're trying to solve to optimize their use. You have to know
where they work and where they don't. At Two Sigma, we do a lot of research and experimentation to
ensure we're applying the tools as effectively as they can be so that we make the best investment
decisions. It's important to separate AI hype from AI reality, and our decades of experience
in innovation with machine learning and artificial intelligence makes it possible for us to make that
distinction. NVIDIA is a valued partner for Two Sigma. We've co-innovated and co-created as partners
over a long journey together. Whether on-prem or in the cloud, we use both hardware and software
to run our AI computations. For example, we use NVIDIA to run the more than 1.7 million
computational nodes 24-7, and NVIDIA's product investment has helped us to attain our own
strategic vision.
