Unlike general artificial intelligence (AI) like ChatGPT from OpenAI, the key to industry-optimized ‘vertical AI’ is securing specialized data in that field. Close collaboration with industry experts is essential for this. The startup ‘LinqAlpha,’ which develops AI specialized for the financial market, is building customized vertical AI by recruiting analysts from global hedge funds, specifically Ha Jaeseon, a specialist in applied AI from Goldman Sachs.
Founded in 2022, LinqAlpha is a startup established by CEO Choi Chan-Yeol, who completed his doctorate in electrical and computer engineering at the Massachusetts Institute of Technology (MIT). LinqAlpha is developing an ‘AI agent’ that helps analysts, who need to analyze hundreds of corporations daily, easily find the information they want by aggregating not just disclosure materials for hedge funds but also social network service (SNS) remarks from chief executive officers (CEOs). Amid high industry interest, LinqAlpha’s client base has nearly doubled in just six months.
Researcher Ha, after working at Hana Securities and then as a TMT (Technology, Media, Telecommunications) analyst at Goldman Sachs, joined LinqAlpha in June of last year. Ha witnessed the trend of AI shaking the entire global industry and judged that AI could be a practical help in the analyst’s job, which requires handling vast amounts of information, so she moved to a startup. As an economics major, participating in building an AI agent was no easy challenge. ChosunBiz met with Researcher Ha on the 20th of last month to hear her detailed story.
― What made you switch to an AI startup?
I have had several conversations with senior researchers from OpenAI who developed ChatGPT while working as an analyst. Through this process, I developed the perception that AI is a technology that can change industrial infrastructure beyond simple automation.
I also experienced that gathering information took a lot of time while performing corporate analysis at my previous job, and inaccuracies frequently occurred during the process of compiling and comparing vast amounts of data. I thought if AI could take over simple tasks, human analysts could focus on thinking and judgment based on the data.
In light of my desire to work in AI, I encountered LinqAlpha and saw the potential to fundamentally change the inefficient process of existing investment research.
― What tasks are you responsible for at LinqAlpha?
I am in charge of planning products based on large language models (LLM). I collaborate with engineers and designers to implement AI agents and coordinate the entire process from quality control to post-launch improvements.
More specifically, I am breaking down the stock research process to determine which aspects can be automated with AI and designing processes to ensure that LLM can accurately interpret financial information.
Recently, I have been focusing on enhancing the ‘Screener’ product. This product helps identify promising investment targets by applying both quantitative criteria like financial statements and qualitative criteria like CEO changes and efficiency improvements. In simple terms, you could say I am working on selecting and automating the quantitative and qualitative criteria needed by investors.
― How has your experience as an analyst helped you?
Having a precise understanding of the stock research process and the ‘pain points’ analysts experience has been a significant help throughout my work. Analysts collect and process vast amounts of information, then interpret it to derive insights.
In that process, errors often occur, and verifying the authenticity of each piece of information is also the responsibility of the analysts. Having personally experienced this reality, I was able to design the requirements and quality control system for the AI agent products to fit the practical work.
― I hear providing information on small and medium-sized enterprises is a strength of yours.
Large companies like Nvidia capture market attention, leading to a flood of data daily. In contrast, small and mid-sized companies suffer from a lack of disclosure materials, and even securities reports are scarce, making analysis more challenging.
LinqAlpha’s paid service collects and processes vast amounts of financial information related to small and mid-sized companies scattered across the web, automatically curating only the essential information that investors need to resolve the problem of information asymmetry. In fact, many hedge funds are showing significant interest not only in large corporations but also in small and medium-sized businesses.
― How much has LinqAlpha grown so far?
LinqAlpha is providing insights through advanced AI technology that analyzes data like a skilled analyst, going beyond simple search capabilities. LinqAlpha has segmented the complex investment research process and offers over ten specialized services for each step.
Personally, I find it significant that globally renowned hedge funds like Schonfeld and Millennium Management, which were major clients of Goldman Sachs, have become LinqAlpha’s clients. Given that AI solutions specialized for the financial sector were not adequately available until now, many financial institutions are showing interest in LinqAlpha. Currently, around 170 financial institutions, including investment banks ranked among Fortune’s top 50 companies, are utilizing the LinqAlpha platform.
― Could AI replace human analysts?
AI will certainly take over repetitive and laborious tasks from human analysts, and through higher-level analytical capabilities, it will act as a partner to analysts in capturing investment opportunities and making decisions.
However, I believe the ability to read the meaning behind the numbers and to detect unspoken trends to make better decisions remains a unique domain of humans.