TL;DR

KelAI is an AI research engine for hedge funds and institutional investors. It captures the full investment research process in one platform and runs it autonomously, from idea generation to data analysis, backtesting, validation, monitoring, and PM feedback.

I spent six years at WorldQuant managing a large systematic equity book, and before that led machine learning at Millennium Management. I saw firsthand how hedge fund research is constrained by human throughput and lost context. Teams can only explore so many ideas, and much of what they learn gets scattered across siloed systems. I am building KelAI because I lived this problem.

KelAI uses AI agents to turn research into a continuous loop, while keeping every new idea grounded in the fund’s knowledge base.

The Problem

Hedge funds scale with research capacity, but scaling with headcount is expensive and hard.

Every PM wants more ideas tested, more datasets explored, more hypotheses validated, and more live signals monitored. But even great teams are limited by human bandwidth.

At the same time, the knowledge they gather is fragmented:

A PM has an idea in a meeting.

A researcher backtests it in their research environment.

A test fails for a useful reason.

A signal works in one universe but not another.

A live model decays.

A risk constraint evolves.

A CIO gives feedback.

These are all valuable sources of information, but they live across emails, notebooks, chats, dashboards, backtests, meetings, and memory. Most of that context gets lost. Teams repeat work, miss old lessons, and lose institutional knowledge when people move on.

The Solution

KelAI unifies the research loop into one continuous workflow, empowering institutional investors with an AI research team they can scale on demand, while capturing the full context so every cycle improves the next one.

It connects to a fund’s data, mandate, universe, risk rules, and research history.
Then AI agents run the full research process autonomously:

- Generate trading signal ideas.

- Write and test research code.

- Analyze market, fundamental, proprietary, and alternative data.

- Run backtests and validation.

- Track why ideas worked or failed.

- Monitor live signal performance.

- Learn from PM feedback and prior research.

KelAI does not remove humans from the loop. PMs still manage the portfolio and own the decisions. 

Why Now

AI agents can now do more than answer questions. They can write code, query data, run long workflows, test hypotheses, and keep context across repeated research cycles.

That matters in hedge funds because alpha is not just one idea. It is the process of finding, testing, improving, rejecting, and combining many ideas over time.

KelAI turns that process into software.

Why Us

The hard part is not saying "AI for hedge funds." The hard part is knowing how hedge fund research really works.

I built KelAI from experience. I managed a large systematic equity book at WorldQuant, worked on machine learning at Millennium Management and worked on data engineering at SolveBio.
I have spent my career building systems that turn data into decisions.

KelAI is designed around the real workflows of institutional investors: data rights, IP boundaries, risk limits, research output, reporting language, and portfolio manager preferences.

Traction

KelAI has been deployed with an institutional investor, and its signals have been running since October 2025, demonstrating that it delivers valuable alpha in production. KelAI also has commitments from prominent institutional players and is in active conversations with several large hedge funds and financial institutions.

Ask

If you run capital, research, data, or innovation at a hedge fund or institutional investment firm, I would love to talk.

We are also hiring senior ML engineers and quant engineers who want to build the autonomous research engine for markets.

Email us at [email protected].

For job applications visit https://kelaitech.com/careers