The role AI just made possible: five quant functions, merged into one person.
A paid four week accelerator for becoming a Quant Merger: one person able to take a quantitative idea from hypothesis to research, code, validation, risk, execution and portfolio judgment, using AI to absorb mechanical work while keeping ownership of the decisions that matter.
The curriculum is revised several times a year as the role moves. Enrol once and you keep every future revision: your access does not expire, and neither does the material.
First cohort applicants join at a founding discount and are added to the Quant Merger Community: access to the SQV3 Research Network and its recruitment reach, plus every future curriculum revision at no extra cost.
Formulates the question.
Tests it against data.
Builds it into working code.
Executes it.
Decides whether it fits the broader portfolio.
Every handoff between those people is a place where context gets lost. AI collapses the need for five separate people: it does not collapse the need for the judgment that used to travel between them.
A Quant Merger is the person who holds that judgment continuously, across the whole lifecycle, using AI to execute the mechanics underneath.
It is not a quant who uses AI tools. It is a restructured way of thinking: a specific discipline about what to delegate, what to never delegate, and how to stay accountable for a model's real world behaviour rather than its backtest.
The framework at the centre of the program. Three tests a task has to pass before AI is allowed to own it.
If it goes wrong, can you undo it. Delegate freely where the cost of being wrong is a rerun, never where it is a filled order.
Can you check the output in the time you actually have. Work you cannot verify before you need it is work you have not delegated: you have simply stopped looking.
Someone has to be able to defend every assumption in the model. That someone is you, whatever wrote the code.
MIT Media Lab researchers gave the failure mode a name: cognitive debt. Measured with EEG, participants who wrote with an LLM showed the weakest neural connectivity of any group, recalled little of what they had produced, and reported a diminished sense of ownership over it. The effect persisted once the tool was taken away.
In quant work that debt has a price attached. A model you did not think through is a position you cannot defend. The three gates exist so the leverage compounds and the judgment does not atrophy.
Read the MIT study↗Every participant carries one quantitative idea from hypothesis to research, code, validation, risk, execution logic and public proof of work. The program is not five disconnected subjects. It is one operating system for modern quant work.
How to formulate a question worth testing, distinguish economic intuition from statistical noise, and decide what AI may own. Reversibility, time gated verifiability and model ownership become the operating rules for every later stage.
Data acquisition, cleaning, exploratory analysis, feature construction, backtest architecture, testing AI generated code, reproducibility, lookahead bias, survivorship bias, transaction costs and execution assumptions. Agentic tools are used as infrastructure rather than as substitutes for understanding.
Walk forward analysis, parameter sensitivity, regime dependence, bootstrap and Monte Carlo thinking, turnover, capacity, slippage, drawdowns, tail behaviour, factor exposure, position sizing and portfolio fit. The standard is not whether a backtest looks good. It is whether you can explain why the edge should exist and what would make you stop trusting it.
How strategies behave once they leave the notebook. Execution logic, monitoring, model drift, data drift, failure conditions and when to retrain or kill a strategy, together with the crypto plumbing a modern quant needs to understand.
Constant product, concentrated liquidity and LP risk.
Perpetuals, funding, options, vaults and protocol mechanics.
Rates, collateral, liquidations and decomposing yield into risk.
Research communication, code and portfolio presentation, LinkedIn positioning and how to turn technical work into something recruiters, funds, desks and protocols can actually assess. The work remains the asset. Visibility makes it legible to the market.
The Quant Merger Community is capped at 200 people. Entry is by application, not by follower count or first come first served access. Everyone fills in the application form and is reviewed before joining.
The goal is bigger than the accelerator itself: to build the SQV3 Research Network, a concentrated group of quants, researchers, traders, developers and technical operators whose work can be trusted, challenged, combined and eventually deployed across real research mandates.
The accelerator is one way into that network. Your research, code, judgment and contribution are what keep you relevant inside it.
Apply to JoinThe community is the starting point for the SQV3 Research Network: a trusted layer connecting strong quantitative talent with the research problems, desks, funds and protocols that need it. It is not a job board. It is a working network built around demonstrated research quality and judgment.
Strong work can lead to collaboration, introductions, research mandates and relevant openings. The point is not simply to place people into jobs, but to create a research network SQV3 can draw on for real quantitative work.
Adapting an existing workflow to AI without losing the rigour that made it work.
Who want to operate across the full stack rather than one function of it.
Bringing the method with them and learning where the market breaks it.
Moving into quant work from another quantitative field.
People who want to become complete quantitative thinkers rather than being trapped inside one quant function.
Also available as a private program for trading desks, funds and research teams: the same material, run against your own book, your own data and your own constraints, so the team leaves with one shared method instead of five private ones.
Scoped in conversation, not self serve.
Bring this to your teamApply to join the first cohort and the Quant Merger Community. Applications are reviewed, and the community is capped at 200 people.
As the first cohort, you join at a founding rate and get extra materials added as the curriculum is built out — a standing thank-you for enrolling before the track record does.