TheBioMLClinic

TheBioMLClinic

TheBioMLClinic

Program Building

Program Building

Program Building

Building an AAV-ML capability is a once-per-organization decision. It can take +2 years to establish. TheBioMLClinic gives your program a shortcut without taking chances.



I built the ML capability at the Broad Institute's Vector Engineering Program from the ground up from first principles, with no existing template. The result is a functioning biotech-grade program, a named patent portfolio, and ML capabilities that have since become the engine for continuous discoveries.

Now I can do it again but much faster so your AAV program does not take chances or wait years to experience the leverage ML can bring.

What this engagement is

I function as the architect of your AAV-ML capability: I design the program structure, define the capability roadmap, identify the right team profile for execution, and validate early milestones.

The deliverable is a permanent internal capability your team owns and operates.

What I bring to the engagement is the judgment that comes from having done this; not the data, models, or institutional work that came out of prior programs. Every program I build is designed around the client's biology, their data, and their path to IND. Nothing is carried over from prior institutional work.

This is a rare, high-stakes engagement. It is structured around milestones, not hours. It is appropriate for organizations that have committed biology infrastructure, a realistic timeline to ML integration, and leadership that understands what a serious computational capability requires to succeed.

If that is not where you are yet, Advisory & Consultation is the right starting point.

What this engagement covers

Experimental design, library strategy, model architecture, candidate prioritization, team orientation on working effectively with AAV-ML, and computational decision support across the life of the program. The scope is defined by your starting point and your target capability, not by a fixed package.

This is the right engagement if —

  • You are launching or already have an AAV program and ML is not yet part of it. And you appreciate the transformative value ML can inject into your program.

  • You need to own the ML capability to maximize its utility across all your AAV work unconditionally.

  • You do not want to keep depending on external partners' capabilities, capacities, timelines, and what you can afford from their catalogue.

  • You are launching or already have an AAV program and ML is not yet part of it. And you appreciate the transformative value ML can inject into your program.

  • You need to own the ML capability to maximize its utility across all your AAV work unconditionally.

  • You do not want to keep depending on external partners' capabilities, capacities, timelines, and what you can afford from their catalogue.

A note for programs with more ambitious design requirements

Some programs are not starting from the standard design space. They are building toward insertion lengths, mutational burdens, multi-loop engineering or translatability constraints that push past what conventional AAV-ML frameworks can address. If that describes your program, the engagement is structured differently: that conversation begins with the problem, not the scope.
[→ Frontier Engagements]

Some programs are not starting from the standard design space. They are building toward insertion lengths, mutational burdens, multi-loop engineering or translatability constraints that push past what conventional AAV-ML frameworks can address. If that describes your program, the engagement is structured differently: that conversation begins with the problem, not the scope.
[→ Frontier Engagements]

Milestone-based. Scoped to your starting point and target capability.

The first conversation is about fit, not about selling.


— TheBioMLClinic


If your AAV program needs to experience the leverage ML can bring without waiting two years or taking chances, the conversation is open.

The first conversation is about fit, not about selling.


— TheBioMLClinic