June 24, 2024
•
5 min read
Get to Know Gradient: Mark Kim-Huang, Co-Founder & Chief Architect
Where were you before Gradient?
“I spent half of my career in financial services, at the intersection of AI and finance. Primarily, combining large-scale data and then applying algorithms on top of that for signal processing. I started in electronic trading and moved up to a quantitative analyst in a statistical arbitrage hedge fund.”
“In the later half of my career, I focused primarily on search and deep learning in the technology sector. I was driven to start Gradient to solve a lot of the problems that I was seeing every day.”
Can you share some of the problems you experienced that led you to start Gradient?
“What I realized throughout my career is that the financial services industry is great at creating infrastructure for data to flow and transactions to occur. For example, just think about all the highly efficient marketplaces or exchanges that exist in financial services.
"However, the industry is not great at consolidating fragmented pieces of data and systems together and then operationalizing them to develop personalized, seamless client experiences that can scale their businesses. This is why I started Gradient."
At Gradient, we specialize in harnessing institutional knowledge to improve tasks at the highest quality possible.
In your opinion, what is the state of AI adoption in today’s market?
“The general observation from enterprise AI startups is that we’re just so early in the curve of AI adoption. The potential of AI is immense. But, lots of enterprises are still catching up with the digital transformation era. They’re still focused on digitizing their processes and transitioning from paper to electronic systems.
From my conversations, the companies that have shown a lot of promise are the ones trying to answer these questions:
How do we become more efficient?
How do we transform ourselves into a more nimble company to grow our market share?
Companies asking these questions will likely thrive in the AI era."
What’s a common misconception about leveraging AI?
“I’d say the most common misconception is that you need an immense amount of data to benefit from AI. Many companies over-emphasize data collection and under-emphasize the capabilities of AI solutions to actually solve their business problems. ”
“This misconception is especially common In financial services. Lots of companies I speak with want to just throw data at a model. But, that’s not always the best solution. Instead, it’s more about defining the problem, optimizing data, and then representing data in a pipeline so that the system can learn and achieve the task that you intend it to do."
“In traditional machine learning, having lots of data was more important so that you could train the model and allow for the capacity of the model (governed by the number of parameters) to fit your task.”
"But, in the era of large language models, you have an infinite capacity in the model. Now, it’s about selecting and optimizing the right samples and then exposing these samples to the task of learning what you want the model to do."
What should finance companies ask when signing an AI provider?
“One piece of advice for companies looking to bring on an AI provider is to be suspicious of fenders who say they can handle any type of data. This is misleading because types of financial data…10k filings… prospectuses...are all very different.
Additionally, it’s important to find an AI provider that can easily align on the success criteria of what you’re trying to achieve.
Can you explain a bit about Gradient’s capabilities? Specifically, how does Gradient’s end-to-end operating system compare to point solutions?
“Many financial services firms rely on AI point solutions to extract and clean data from documents. Yet, a challenge these companies often run into months later is realizing these tools stop short of solving the whole problem. They don’t address the critical steps of normalizing and reasoning with data to deliver real business outcomes.
“Understanding these common challenges, Gradient built a complete end-to-end platform that goes beyond surface-level AI applications. Gradient offers normalization & inference capabilities to enable organizations to reason with data and deliver actionable business results.”
What is Gradient in its simplest terms?
“We are the AI operating system for financial institutions. We allow financial institutions to bring in their business systems and data and help them automate and operationalize business processes that are specific to their needs. We do all of this on one platform that interoperates and harnesses data from their organization.”
Thanks for takin the time to join us!
We hope that you’ve enjoyed this episode of Get To Know Gradient as we interviewed Co-Founder & Chef Architect, Mark Kim-Huang.
Interested in speaking with Mark about how a high-performing, cost-effective custom AI system could benefit your financial services business?
Contact the Gradient team today to learn more.
Share
