Bhavdeep Sethi is the founding engineer at Frec, former Twitter adtech lead and a contributor to the Forbes Technology Council. He’s paving the way for the next generation of passive investing, making AI and machine learning-driven strategies accessible through Frec’s first-to-market multi-index-capable platform. Today, he joins us to share his insider perspective on the exciting new opportunities available to retail investors, the broader landscape of fintech, and why AI’s emergence may signal a shift away from traditional ETFs.
Thank you for taking the time, Mr. Sethi. Could you tell us what led you to specialize in fintech? How does someone find themselves the architect behind Frec’s direct indexing algorithm?
Thank you for having me. My career has always been rooted in tech and data-driven systems. Before Frec, I was a Senior Software Engineer at Twitter, where I focused on optimizing real-time ad placements—a role that naturally led me to apply machine learning techniques in the financial sector, and to the much livelier stock market. What we’ve built at Frec is really a culmination of those experiences. It’s given me the opportunity to apply ML in a very novel way, especially since direct indexing had largely fallen off the radar for non-institutional investors. When you dig into the nuances of real-time tax lots, you see just how much value is being left on the table—just how much tax optimization can be achieved with the right tools. At Frec, we want to make sophisticated investment strategies accessible, and I saw a tremendous opportunity with direct indexing to do just that.
Frec has recently reached a significant milestone—$100 million in assets within nine months of launch. Could you walk us through the direct indexing algorithm you helped develop and explain how it sets Frec apart in the market?
Certainly. Frec’s direct indexing algorithm is designed to offer retail investors an unprecedented level of tax efficiency, which was traditionally only available to institutional or high-net-worth investment strategies.
Direct indexing enables investors to directly own the underlying securities of an index, rather than investing through a pooled vehicle like an ETF. This approach allows for granular tax-loss harvesting (TLH) on individual components of the index. Under normal circumstances, executing these transactions might require a dedicated advisor with sector-specific knowledge, just to avoid wash sales.
The algorithm we developed at Frec leverages real-time data to assess the cost basis of each security and executes trades that optimize tax outcomes. This is bookkeeping efficiency at its best, and happening at very regular intervals, automatically. And by enabling daily TLH, we’re able to drive much greater tax savings for our investors. Combine that with our industry-first ability to harvest losses across multiple indices, it’s not surprising that Frec is gaining traction with investors. They’re recognizing the opportunity and the value. We’re continuing to explore the deep learning capabilities provided by AI to predict potential wash sales and adjust trading strategies, while ensuring compliance with IRS regulations—there are some promising early results.
Frec is really pushing the boundaries of what’s possible for retail investors. How do you see AI and machine learning changing the broader fintech landscape?
AI and machine learning have always been, and will continue to be, integral to fintech. These technologies are purpose-built to process vast amounts of data, and make decisions at scale and speed far beyond human capability, all while continuously learning and improving. I’ve had the rare opportunity to help make these technologies more accessible and effective for retail investors—bringing the quantitative capabilities of institutional players into the hands of everyday people.
One of the most exciting developments is how the type of data AI consumes is evolving. We’re seeing a rise in natural language processing (NLP) for handling unstructured data, such as news articles and social media. This allows investment strategies to be more responsive to market sentiment and make predictive adjustments. Of course, like any data-driven tech, it’s only as good as the data it processes—garbage in, garbage out—but we’re seeing promising exceptions that are overcoming the odds.
Beyond predictive analytics, AI’s growing capability to deliver hyper-personalized investment strategies is really fascinating. We’re already seeing models that can tailor portfolios to individual preferences, mixing and weighting factors like risk tolerance, investment horizon, and ethical considerations like ESG. As these products become more refined, they’ll offer even more precise adjustments that respond to changes in an investor’s life circumstances or broader market shifts. This level of personalization was unimaginable a few years ago, and I’m proud to be at the forefront of these changes.
For those studying AI and machine learning, do you see fintech as a promising field to apply their skills?
Absolutely. Fintech is where some of the most exciting AI/ML innovations are happening. The demand for expertise is growing rapidly because financial advisors aren’t simply picking up neural networks and plugging them in—it’s become the new face of quant trading. The problems we’re tackling are complex and have a wide-reaching impact, given how interconnected global markets have become. It’s a space where you can really, actively play a part in shaping the future of personal wealth.
Thank you again, Mr. Sethi. Readers can find your writing on Forbes and Hackernoon. For those interested in more ‘inside baseball,’ Frec’s white papers provide a wealth of investment knowledge and offer a glimpse at where AI-driven fintech is headed next, including a detailed look at Mr. Sethi’s algorithm.
