00:00:00–00:40:00
Gerardo Lemus, Head of Data Science for Lantern AI, shared his background and experiences leading him to join LanternAI in this episode of “The Python People Podcast.” Previously, he studied electrical engineering and computer science at MIT and developed an interest in statistical learning. He then transitioned from finance into fundamental portfolio management before eventually returning to his academic roots to learn about the latest machine learning theories and techniques around 2015. In discussing the factors that led to the rise of machine learning systems, Lemus highlighted the availability of data and computing power as game changers. He detailed his journey from a clinical background to relearn the latest techniques and gain experience in startups and consulting, aiming to help financial institutions use these techniques. Lemus then discussed the inception of LanternAI, which was started in February 2022 to provide investors with accessible and actionable financial data. The team created a portfolio monitoring tool and aimed to become a source of alternative data points for clients. However, developing a secure platform for the private equity markets proved to be a significant challenge due to data security concerns. Lemus expressed his excitement about the potential of LanternAI to revolutionize data science by enabling business people to connect directly to data and increasing productivity by a factor of ten. Despite the limitations of generative AI in creating new ideas, he believed that LanternAI’s ability to translate between domains would become increasingly powerful over time. Throughout the interview, Lemus emphasized the importance of a comprehensive understanding of the industry and competitors when starting a business, using his previous experiences as examples. He concluded the interview with the advice he received from his father: “It’s better to regret having done something than to regret not having done it
See less
- 00:00:00 In this section of “The Python People Podcast,” Gerardo Lemus, Head of Data Science for Lantern AI, shares his background and how he got into tech and data. While studying electrical engineering and computer science at MIT, he spent time as a research assistant at the AI lab and developed an interest in statistical learning. After completing his PhD, he moved into the finance industry as a quantitative analyst and strategist, eventually transitioning into fundamental portfolio management. Around 2015, he grew restless, noticing the renewed interest in AI. Having previously believed that areas like finance would remain human domains, he was excited about the advances in AI and decided to return to his academic roots to learn about the latest theories and techniques
- 00:05:00 In this section of the podcast, Gerardo Lemus from LanternAI discusses the key factors that led to the rise of machine learning systems that outperformed traditional statistical techniques. He mentions that while the theories and techniques remained common, the availability of data and computing power were the game changers. He shares an example of an academic project that gathered millions of images and had humans manually tag them, which enabled the training of modern systems. Lemus also talks about his personal journey from a clinical background to relearn the latest techniques, and how he gained experience in startups and consulting to help financial institutions use these techniques for their purposes
- 00:10:00 In this section of the podcast, Gerardo Lemus discusses his past work experience in consulting for hedge funds and banks, where he focused on making financial data accessible and actionable for investors. He highlights the importance of understanding the industry and the investors’ needs to provide valuable insights.
- 00:15:00 In this section of “The Python People Podcast” with Gerardo Lemus of LanternAI, the discussion revolves around the creation of the company and their initial goals. LanternAI was started around February 2022 after a period of brainstorming, with the name already chosen. The team aimed for augmented intelligence, enhancing human abilities rather than creating an artificial general intelligence. They operated in stealth mode initially and focused on creating specific products for clients, primarily fund managers. Their first product was a portfolio monitoring tool, designed to provide clients with user-friendly access to their existing data. The team also started exploring advanced AI commercial models to enhance their systems.
- 00:20:00 In this section of “The Python People Podcast,” Gerardo Lemus from LanternAI discusses the challenges of building a secure platform for private equity markets. The main hurdle is ensuring data security, which limits the use of certain tools due to their inability to provide completely secure environments.
- 00:25:00 In this section of the “Python People Podcast” with Gerardo Lemus from LanternAI, the discussion revolves around the challenges of working with various data sources and ensuring data security and privacy. Lemus explains that the greatest challenge is not in using specific tools but in ensuring the legality and safety of data sources. The company takes robust measures to vet and secure all data, prioritizing lawful processes and avoiding copyright issues. Only then can they consider which tools are suitable in a secure development environment. Lemus expresses excitement about the evolution of data science and AI, specifically the ability of large language models to translate between natural language and code. This could potentially eliminate the bottleneck of data scientists converting natural language queries into SQL queries and increase productivity by allowing the business side to ask questions directly to the data
- 00:30:00 In this section of the podcast, Gerardo Lemus discusses LanternAI’s potential to revolutionize data science by allowing business people to connect directly to data and increasing productivity by a factor of 10. He highlights the importance of accurate translation between domains and acknowledges the limitations of generative AI in creating new ideas. Lemus is excited about AI’s ability to translate between domains and believes it will become increasingly powerful over time. As a tech and data leader, he emphasizes the importance of understanding the business and focusing on solutions that increase profits for the company
- 00:35:00 In this section of “The Python People Podcast,” Gerardo Lemus discusses the importance of having a comprehensive understanding of the industry and competitors when starting a business. He emphasizes that having competitors signals that there is a market for the product, and if you can improve on what they are offering, you will deliver the best service to your companies. Lemus also shares his experience and observance that people advancing into more senior positions possess an overall business perspective and a pragmatic approach to choosing the right tools for the job. The podcast concludes with Lemus sharing the advice he received from his father: “It’s better to regret having done something than to regret not having done it.” This philosophy encourages people to act on their ideas, learn from their failures, and not live with the regret of inaction
- 00:40:00 In this section of the “The Python People Podcast” interview with Gerardo Lemus of LanternAI, the interviewer expresses confidence in Lemus’ future success. Lemus responding graciously and expressing his appreciation for the support. The conversation concludes with the interviewer wishing Lemus well and signing off
