Creative LLM use cases - Job post analysis, say-do scores, merger arb, expert network MCPs
Plus: Guides on Claude Code, Claude Cowork, Claude for Excel and more
GenAI job post analysis
We’ve been collecting hedge fund GenAI job posts over the past month to identify creative LLM use cases, and thought we would analyze them to see what else we could learn:
Funds with the most job listings
Most common technologies
Key takeaways from the data
Average annual salary in $212.5K, with a range from $150K to $300K
Teams prefer vendors for models and storage, open source for everything else
OpenAI mentioned twice as much as Anthropic
AWS is the most popular hyperscaler
While Balyasny, Millennium, Point72 dominate hiring, we did not find evidence of open fundamental-focused GenAI roles at Citadel
If you’d like the full dataset, please reach out
Creative LLM use cases
Bloomberg Businessweek used Claude to review 1,500 hours of livestream footage of influencers playing Stake, a crypto gambling site, to reveal the company was rigging bets (Bloomberg, thanks to Byrne Hobart)
Reporters used Anthropic’s Claude, a large language model, to analyze footage frame by frame and determine the balance, bet and games being played during livestreams
An investor vibe coded an LLM that identifies past CEO claims and whether they bore out into a “say-do” score for every management team (@dahu7744)
Good thread on the process of iterating with Claude Code until it can correctly generate excel models (@tomasrice_au). Related, OpenAI launched ChatGPT for Excel (OpenAI)
Guide to using Claude Code / Cowork for investment research by CEO of Daloopa (@oneThomasli)
JPAM hiring data scientist to analyze sellside notes and news to identify trending and emerging themes (LinkedIn)
Case study on Jefferies equity research internal alternative data LLM chat (Databricks)
This multi-source response surfaces analytical angles that analysts may not have explicitly requested, enabling corroboration across independent sources.
Roundup of internal LLM tools at BAM, Citadel, Point72, etc (@TheValueist)
Merger arb
A merger-arb ETF manager describes his LLM system (@JulianKlymochko):
For example, we have Equity Research Analyst agent that writes an initiating coverage report on the merger target. Next, we have our Legal M&A Analyst agent, that summarizes merger agreements and proxy statements, and our Antitrust Analyst agent, that analyzes market shares along with the DOJ / FTC, EC, China SAMR, and other global regulators would view the deal, in addition to ascribing probabilities of antitrust clearance / 2nd requests / merger challenges.
How Balyasny built its merger arb bot (OpenAI):
early feedback from merger arbitrage teams revealed that agents needed to continuously re-evaluate deal probabilities as new filings or press releases came in. The Balyasny team quickly extended agent planning capabilities and tool access, replacing a slow, manual workflow with real-time probabilistic monitoring
Expert network MCP
Third Bridge launched an MCP for its transcript library (press release). AlphaSense/Tegus offers one as well but GLG and Guidepoint currently do not have public MCP or API endpoints.
My take: seems like all the major transcript libraries will offer MCP / API access soon. Demand for transcript content should meaningfully increase as agents aren’t limited by cognitive ability to process transcripts. However, I expect pricing to fall even more as the search cost across providers will go down. Networks will compete on their ability to access experts exclusively - and unique experts will benefit accordingly. I also wouldn’t be surprised to see consolidation.
Follow for more investor LLM workflows
If you would like to discuss incorporating LLMs into your research process, reply to this email or reach out via X or LinkedIn.