Beyond the AI Hype with Jason Strimpel

In this episode of The Hedgineer Podcast, host Michael Watson sits down with Jason Strimpel, founder of PyQuant News, long-time Pythonista, and AI enthusiast.

They dive deep into the practical and philosophical implications of artificial intelligence in both asset management and daily life. They break down how the agentic loop works with the Anthropic Agent SDK, the components of good evaluation frameworks, and even how to talk to your kids about AI 

Key topics covered in this episode:

• Drugs, Sex, and AI: Jason shares his thoughts for new parents: the three things you now need to talk to your kids about are drugs, sex, and AI. Michael and Jason then discuss the difficulty of explaining the difference between humans and AI-powered avatars or toys to children.

• The "Agentic Loop": The discussion breaks down the simplicity and power of the agentic loop, identifying it as the "core substrate" of modern agentic frameworks. This framework allows a language model to loop, use tools, and determine when to exit to solve complex problems.

• "Evals are the Product": Michael and Jason iterate on the concept that a robust set of evaluations is the real product. If you can use evals to demonstrate that an agent has harnessed intelligence to solve a specific problem space, you "own that problem".

• AI vs. Python's Rise: They draw parallels between the current AI boom and the rise of Python in the early 2010s. Both technologies were initially met with skepticism for being "black box" interpreted systems, yet they unlocked massive productivity boosts.

• Capital Allocation and Moats: The conversation tackles the modern challenge of allocating capital and defending a "software moat" when new AI tools and infrastructure are being commoditized by hyperscalers at an incredible speed.

• The PyQuant News Story: Jason shares the origin story of his popular PyQuant newsletter, which started as a personal WordPress site for bookmarking research papers and grew into a major resource for the quantitative finance community.

Hosted by Michael Watson, The Hedgineer Podcast dives into AI technology and data in the hedge fund, asset management, and prop trading space.

Follow The Hedgineer Podcast:

YouTube: (https://www.youtube.com/@hedgineer)

LinkedIn: (https://www.linkedin.com/company/90976838)

Twitter: (https://x.com/hedgineering)

Instagram: (https://www.instagram.com/hedgineer/)

Don't forget to like, subscribe, and hit the notification bell to stay updated on our latest episodes!

Hedgineer.io

Hosted on Acast. See acast.com/privacy for more information.

Kuzu, Knowledge Graphs, and the AI Revolution with Prashanth Rao

In this episode of the Hedgineer Podcast, host Michael Watson is joined by Prashanth Rao, AI Engineer at Kuzu, for a deep dive into the world of embedded graph databases and their pivotal role in the age of AI. Building on the previous week's discussion of the columnar database DuckDB, this conversation explores Kuzu, a parallel concept focused on creating high-performance, embedded graph databases.

Michael and Prashanth explore why Kuzu's unique architecture—an embedded, columnar, and strictly-typed system—is delivering incredible speed and scalability for complex analytical queries. They discuss the resurgence of interest in knowledge graphs, driven by the need to impose structure on data for modern AI and LLM workflows. Prashanth explains how LLMs are revolutionizing both the upstream construction of graphs from unstructured data and the downstream querying of these graphs by translating natural language into the Cypher query language.

Tune in to learn about:

• What Kuzu Is: An in-process, embedded graph database designed from the ground up for query speed and scalability, blending the benefits of columnar processing with the property graph data model.
• AI and Graph Synergy: How large language models (LLMs) assist in both building knowledge graphs through structured output extraction and accessing them via natural language to Cypher translation.
• Modern Data Stack Integration: The role of Kuzu as a powerful secondary semantic layer that interoperates seamlessly with primary data stores in data lakes (like Parquet files on S3), DuckDB, and Postgres.
• Programming with LLMs using DSPy: A detailed look into using DSPy to create structured, modular, and optimizable prompts for more reliable LLM applications, bridging the gap between deterministic code and fully agentic workflows.
• The Enterprise Knowledge Graph: A discussion on how a knowledge graph can become the single source of truth for understanding complex organizational workflows, data lineage, and interdependencies across an entire enterprise.
• The Future of Kuzu: A preview of what's next for the Kuzu project, including new graph algorithm capabilities and enhanced concurrency for reads and writes.

Whether you're a data engineer, AI practitioner, or a leader in the asset management space, this episode provides a masterclass on leveraging modern database technology to build powerful, scalable, and intelligent applications.

Hedgineer.io

Follow The Hedgineer Podcast:

• YouTube: (https://www.youtube.com/@hedgineer)
• LinkedIn: (https://www.linkedin.com/company/90976838)
• Twitter: (https://x.com/hedgineering)
• Instagram: (https://www.instagram.com/hedgineer/)

Hosted on Acast. See acast.com/privacy for more information.

In this episode of The Hedgineer Podcast, host Michael Watson is joined by special guest Rusty Conover, the world's most prolific DuckDB extension builder, for a masterclass on building the next generation of real-time, large-scale data systems.

Rusty, who has an extensive career in data engineering, including at multi-manager hedge funds, pulls back the curtain on what makes DuckDB so revolutionary for developers and data engineers. They explore how its blazingly fast, in-process, C++-based architecture is challenging the big data status quo. The conversation provides a deep dive into the powerful ecosystem growing around DuckDB, from the Apache Arrow columnar format to the evolving landscape of open table formats like Iceberg, Delta Lake, and the new DuckLake.

Join them for a detailed discussion on the nitty-gritty of modern data infrastructure, whether you're building enterprise data platforms or looking for the most efficient tools for your analytics workload.

In this episode, you will learn about:

The DuckDB Revolution: What makes this "blazingly fast" in-process database a game-changer that can simplify and replace entire ETL stacks.

A Tour of DuckDB Extensions: A look inside some of the 15 extensions Rusty has built, from Airport for integrating with Apache Arrow, to Crypto, ShellFS, and TextPlot.

Diving into Apache Arrow: An explanation of the columnar in-memory data format, zero-copy operations, and the Arrow Flight RPC mechanism for efficiently moving data.

The Battle of Open Table Formats: A comparison of Iceberg, Delta Lake, and the new database-centric approach of DuckLake.

DuckDB vs. The World: How DuckDB stacks up against KDB for financial data, ClickHouse for analytics, and its role alongside large-scale compute engines like Apache Spark.

Parquet Deep Dive: The key differences between Parquet V1 and V2 and the importance of modern compression strategies and encodings.

The Future of DuckDB: A sneak peek at powerful upcoming features like time travel and the MERGE INTO statement for simplifying change data capture (CDC) pipelines.

Hosted by Michael Watson, The Hedgineer Podcast dives into AI technology and data in the hedge fund, asset management, and prop trading space.

Follow The Hedgineer Podcast:

YouTube: (https://www.youtube.com/@hedgineer)

LinkedIn: (https://www.linkedin.com/company/90976838)

Twitter: (https://x.com/hedgineering)

Instagram: (https://www.instagram.com/hedgineer/)

Don't forget to like, subscribe, and hit the notification bell to stay updated on our latest episodes!

Hedgineer.io

Hosted on Acast. See acast.com/privacy for more information.

Welcome back to Season 2 of The Hedgineer Podcast, where host Michael Watson talks data, AI, and technology in the hedge fund and asset management space.

In this episode, Michael sits down with Didier Lopes, the founder of OpenBB, for a conversation that started five years ago during the meme stock boom. Michael reached out to Didier in 2021 after he shared his "Gamestonk Terminal" on r/wallstreetbets. Michael reached out to Didier if he was interested in a job on his team at Citadel but Didier had other ideas. That initial LinkedIn message from Michael ended up in the pitch deck that helped Didier raise nearly $9 million for his company and the rest is history.

Didier recounts his journey from being a software engineer who automated his personal investment research to becoming the creator of a massively popular open-source project. The Gamestonk Terminal went viral on WallStreetBets and Hacker News, gaining 4,000 GitHub stars in less than 24 hours by cleverly aggregating the free API tiers from hundreds of different data vendor and has evolved into what OpenBB is today.

Join us as we dive into the evolution of a viral tool into a sophisticated enterprise platform and explore the disruptive forces reshaping the financial data landscape.

In this episode, you will learn about:

• The Origin Story: How a canceled flight during the COVID-19 pandemic led to the creation of the Gamestonk Terminal.
• From CLI to Enterprise: The strategic pivot from a command-line tool for retail investors to a customizable, SDK-first workspace for financial firms (and then possibly back to CLI with agents).
• The Data Vendor's Dilemma: A fascinating discussion on the conflict facing companies like FactSet and Capital IQ, whose UI and terminal businesses are threatened by the rise of AI and direct data access through protocols like MCP (Model Context Protocol).
• The Future of Workflows: Michael explains how he uses AI agents like Claude Code combined with CLI tools to build powerful, autonomous systems for everything from risk management to analyzing entire enterprise data warehouses.
• The Autonomous Firm: A look into a future where AI enables multi-billion dollar asset managers to operate with fewer than ten people.
• The Rise of "Acqui-hires": Why the next wave of M&A might value a team's ability to leverage AI far more than the IP they've built.

This is episode is for anyone at the intersection of finance, data science, and artificial intelligence.

Hedgineer.io

Follow The Hedgineer Podcast:

• YouTube: (https://www.youtube.com/@hedgineer)
• LinkedIn: (https://www.linkedin.com/company/90976838)
• Twitter: (https://x.com/hedgineering)
• Instagram: (https://www.instagram.com/hedgineer/)

Hosted on Acast. See acast.com/privacy for more information.

Welcome back to the Hedgineer Podcast!

We're kicking off Season 2 with one of our favorite guests, Thomas Li, Co-founder and CEO of Daloopa (https://hubs.ly/Q03C7TSR0).

In this episode, we delve into the significant advancements Daloopa has made, including its groundbreaking announcement with Anthropic FIS and its integration into OpenAI. Thomas shares how Daloopa's high-quality, error-free financial data is the key to unlocking the true power of AI for financial analysts. By plugging its verified database into Large Language Models (LLMs), Daloopa is eliminating the critical problem of "hallucination," where AI generates incorrect numbers. This opens up analytical workflows that were previously impossible.

In This Episode, We Discuss:

• Daloopa's core mission: A relentless focus on accuracy (over 99.9%), comprehensiveness, and speed in financial data.
• The "Hallucination Problem": Why LLMs like ChatGPT and Claude often fail at financial analysis by generating wrong numbers from the internet.
• Unlocking New Workflows: How AI with reliable data enables analysts to create deep research reports, summarize 150 earnings calls before the market opens, or run LBO analyses on entire portfolios instantly.
• A Real-World Example: How a PM used the technology to generate a comprehensive report on the impact of tariffs across all reporting companies before the market opened—a task previously considered impossible.
• The Future of Pod Shops: How AI impacts hedge fund strategy by increasing a portfolio manager's "at-bats" and "slugging".
• Behind the Scenes: The story of the Daloopa and Anthropic FIS partnership, rooted in a mutual focus on data quality and solving real customer problems.
• The "Conference Problem": A look into Daloopa's research on creating a true AI agent inside Excel to assist analysts in real-time during management meetings.

About Our Guest: Thomas Li is the Co-founder and CEO of Daloopa. This company has cemented itself as the highest quality data source for company-specific financials and KPIs. Daloopa is trusted by analysts at the world's leading long-short pod shops and multi-manager funds.

Hosted by Michael Watson, The Hedgineer Podcast dives into AI technology and data in the hedge fund, asset management, and prop trading space.

Follow The Hedgineer Podcast:

• YouTube: (https://www.youtube.com/@hedgineer)
• LinkedIn: (https://www.linkedin.com/company/90976838)
• Twitter: (https://x.com/hedgineering)
• Instagram: (https://www.instagram.com/hedgineer/)

Don't forget to like, subscribe, and hit the notification bell to stay updated on our latest episodes!

Hedgineer.io

Hosted on Acast. See acast.com/privacy for more information.

Special Announcement: Kicking Off Season 2!

In this season 2 kickoff episode, host Michael Watson welcomes listeners back for the second season of The Hedgineer Podcast, the go-to source for discussions on AI, data, and technology within asset management, hedge fund, and prop trading space.

Get ready for an exciting new season featuring a lineup of entrepreneurs and thought leaders. The season kicks off with a return guest, Thomas Li, the CEO of Daloopa, to discuss the significant changes in the AI and data landscape since his last appearance.

Michael also shares the evolution of Hedgineer from a passion project podcast into a thriving company that builds data platforms for a growing list of clients. He details how Hedgineer is leveraging AI tools, such as Claude Code, to accelerate client onboarding and software development, emphasizing the increasing importance of expertise in the age of AI.

On a personal note, Michael shares the joys of becoming a father, life in New York City with his daughter, and highlights from a recent family trip to Spain's Basque Country and Bordeaux, France. Join us for a season of insightful conversations, technological deep dives, and personal growth.

We are incredibly excited for the new topics, guests, and conversations we have lined up.

Listen to the full special announcement now and get ready for an awesome new season!

Hosted on Acast. See acast.com/privacy for more information.

Welcome to Episode 16 of The Hedgineer Podcast! In this episode, host Michael Watson dives deep into the world of technology, data, and AI in the asset management, hedge fund, and proprietary trading space. Joining Michael is Christina Qi, the founder of Databento.

Michael kicks off the episode by introducing Christina and discussing the disruptive impact of Databento on the real-time market data experience within the asset management hedge fund space. Christina shares her extensive background, including nearly a decade of running a hedge fund and the challenges she faced due to poor data quality and high costs.

Christina explains the motivation behind founding Databento, driven by the need for better data solutions tailored for machine learning and AI applications in trading. She emphasizes the importance of having real-time, high-quality data delivered via API, which was previously lacking in the market.

Throughout their conversation, Christina and Michael explore various topics, including the scalability of data vendors, the differences between working in a data vendor versus a traditional financial firm, and the challenges of managing data in the finance industry. They also discuss the evolving landscape of investors, highlighting the rise of prosumer-type data consumers who require more advanced data services than typical retail users.

The episode also delves into the impact of community-driven platforms like Reddit and the importance of genuine content in fostering meaningful discussions. Christina shares her experiences with different online communities and the challenges of maintaining authenticity in a space often dominated by advertisements.

Overall, this episode provides valuable insights into the intersection of technology, data, and AI in the finance industry, showcasing the innovative solutions offered by Databento and the broader trends shaping the future of asset management and trading.

Don't miss this insightful conversation! Like, comment, and subscribe for more episodes of The Hedgineer Podcast.

Hedgineer.io

Hosted on Acast. See acast.com/privacy for more information.

Data is maturing at an incredible rate. In turn, the value of data is increasing as new strategies and productization continues to evolve. In episode 15, Eagle Alpha CEO Niall Hurley, discusses the changes and developments in the alternative data space.

Niall highlights emergence of corporates, often times b2b SASS companies, realizing new high margin revenue opportunities by monetizing their data by working with companies like Eagle Alpha to deliver unique insights to their clients. By allowing EA to take on the compliance and economic feasibility testing of bringing new data sets to market, Niall is at the tip of the spear for bringing new insights into the market place faster and helping make better capital allocation decisions.

Historically new data sets are brought to market by an entire entity whose business model is to resell data (Think 1010, Yodlee, etc, or to create new data sets as an aggregator, think Yipit, MScience, etc). It is rare for a corporate to sell direct in a b2c model. It's growing increasingly more challenging to figure out compliance, difficult to find distribution unions, and is not always a significant enough revenue opportunity to justify the resources required to execute.

However, Niall believe change is among us his company is positioned well to act as a market place for the high managing, orthogonal revenue opportunities for corporates to monetize their data.

Additional topics include:

The exploration of new data categories and the continued support for data delivery and go-to-market strategies.

Advancements made by quantitative and multi-strategy funds in utilizing alternative data, and the challenges faced by smaller funds in adopting these data sources

The need for a robust data infrastructure and platform to effectively integrate and analyze alternative data

Is there a need for open-source solutions in risk modeling?.

It’s all here in this incredible episode of the Hedgineer Podcast. Listen here.

Hosted on Acast. See acast.com/privacy for more information.

At Hedgineer, we believe data has a voice and we are committed to helping you hear what its saying.

In this episode, former Goldman Lead Engineer, now Hedgineer Managing Director Jhanvi Virani joins us to discuss how hedge funds can deliver alpha to their investors by leveraging the Hedgineer platform. Gone are the days of siloed data and non existent processes. 

To thrive as a hedge fund in 2024 and beyond, you need to interact with your data to unlock its true value.Jhanvi explains how her skill set and years of experience at Goldman uniquely position her to solve the most challenging data and analytics problems our customers face. She believes in the power of storytelling, that is has a direct impact on the way your investors evaluate their relationship with your firm.

This is a DO NOT MISS episode. Thanks for joining us.

Key points include:

Interconnectivity between Risk, Research, Trading, Operations, and Investor Relations

Breaking down data silos

Unlocking data

Enabling your data to tell stories

Define and unify processes

Bring your own cloud

Hosted on Acast. See acast.com/privacy for more information.

Just about everything in the world of finance is data. It’s only a matter of time that the early adopters and educated users of AI will outrace their competitors to the top.

In Episode 13 of the Hedgineer podcast, host Michael Watson interviews Rob Krzyzanowski, an expert in AI and machine learning with experience at notable firms like Avant, Spring Labs, and Citadel - and also a partner at Hedgineer.

Rob offers his best-in-class view on the value of AI, and specifically, how it translates into platforms that can produce tangible results for hedge funds. We also get a simplified value proposition perspective for large language models, and how they're fundamentally changing the way funds operate.

The conversation also provides a forward-looking view on balancing AI advancements and their application to:

-risk management

-stock classification

-operating efficiencies

-cost reductions

As the team continues it's expansion efforts, Michael wraps the episode with a call for talent. Hedgineer's growth is fueled by world class people. Come join us.

Hedgineer.io

Hosted on Acast. See acast.com/privacy for more information.