Data-Driven Life Sciences · 2 hours of your time, an answer in 3 weeks
Sign up a project

Data-Driven Life Sciences · KTH & SciLifeLab · Autumn 2026

Call for research problems & data

We send a forward-deployed AI scientist to work on your data.

  • 2 hoursyour total time
  • 3 weeksto an answer
  • Freeno obligation

Two 1-hour meetings is the whole ask. In between, someone from the course builds a working analysis on your data with an AI agent and hands back a report and an app you can run yourself.

Every project gets an answer — if nobody is free to take yours, our AI data scientist does
How a forward-deployed AI scientist works on your problem Your question and your data go to a forward-deployed scientist, who interviews you and then directs an AI agent. The agent writes and runs the code — loading data, analysing, plotting, running controls. The scientist checks every number and hands back a report and a runnable app. The cycle repeats until the answer is sharp. YOU the researcher “Does it do anything?” + YOUR DATA INTERVIEW Forward-deployed AI scientist directs & verifies DIRECTS AI agent DOES THE LABOUR load the data write the code run the controls plot the result CHECKED YOUR ANSWER REPORT + APP AND BACK TO YOU — EACH ROUND SHARPER day 0 2 hours of your time ≈ 3 weeks later Schematic. Chart shape is illustrative, not measured data.
  • You Your question & your data “Does it do anything?” — plus the dataset nobody has had time to read.
  • Interviews you, then directs A forward-deployed AI scientist Specifies · directs · verifies. Someone from the course, not a black box.
  • Does the labour An AI agent The person supplies the judgement; the agent does the work. load the datawrite the coderun the controlsplot the result
  • Checked, then handed back Your answer A written report and an app you can run yourself — then round two, sharper.

2 hours of your time · ≈ 3 weeks end to end

Who is asking

We are a course — and we are training
forward-deployed AI scientists

Data-Driven Life Sciences (SK2538 / FSK3538) is a course at KTH, run with SciLifeLab by the AICell Lab for the DDLS community — master's students, PhD students, postdocs and researchers alike. We are its participants and its teaching team.

The course trains one specific role, and it can only be practised on someone else's real problem. That is the entire reason we are writing to you.

What is a forward-deployed AI scientist?

The name is borrowed from software. A forward-deployed engineer — a role made well known by Palantir, and since 2024 used at the major AI labs — is the engineer who is sent to sit with the customer, rather than the one who stays with the codebase. They learn the customer's problem first-hand and build against it directly, in days rather than quarters.

The life-science version is what this course trains:

A scientist who can sit down with someone else's biological problem, interview them until the question is precise, and come back within days with a working, checked answer — with an AI agent doing the labour and themselves supplying the judgement.

What makes it new is the agent. An AI agent can now execute almost any analysis you can specify, which moves the scarce skill from implementation to specification, direction and verification. The failure mode stops being "the code didn't run" and becomes "the code ran beautifully on the wrong question" — so the person's job is to choose the right question, then check the answer hard enough to sign it.

More on the role, and how the course is built around it, further down this page.

Visit the course site — ddls.aicell.io

  • SK2538 / FSK3538MSc, PhD and researcher tracks · 7.5 credits
  • KTH · SciLifeLabrun by the AICell Lab, part of the national DDLS programme
  • Since 2022this is the fifth year the course has run
  • 26 participantsmaster's and PhD students, postdocs, researchers

This is deliberately short

It is not an open-ended collaboration and we are not asking for a grant, a co-authorship or a commitment. Two 1-hour meetings, about three weeks, then it is done — and you keep whatever was built.

The problem we are both in

Most projects die in the gap between two people

One person owns the question. Another owns the methods. They are rarely the same person, and the queue between them is longer than the project's patience.

The ownership gap, and the role that closes it On the left, the scientist who owns the question, the sample and the knowledge of what wrong looks like. On the right, the analyst who owns the methods and the pipelines, booked for four months. Between them a shaded gap labelled where most projects die. Below, a single figure bridges the two, interviewing on one side and directing methods on the other. THE GAP where most projects die The scientist owns the question · owns the sample knows what wrong looks like cannot run the analysis The analyst owns the methods · owns the pipelines has never seen your sample booked for four months The forward-deployed scientist interviews one side · directs the other · one person
The gap closed by one person — the role this course trains.
  • One side The scientist Owns the question and the sample. Knows what wrong looks like — and cannot run the analysis.
  • Between them The gap Where most projects die.
  • The other side The analyst Owns the methods and the pipelines. Has never seen your sample — and is booked for four months.
  • The role this course trains The forward-deployed scientist Interviews one side, directs the other. One person.

What changed

An agent can run every step. It cannot choose the right one.

Answering a real question means climbing down a ladder, from a human sentence to something a machine can run and a human can check. AI agents can now execute every rung of it.

What they cannot do is tell you that rung two was the wrong quantity to measure — that the thing you chose to count was not the thing that matters. That judgement is yours, and it is why this still needs a scientist in the loop.

Which is exactly why our students need your problem: the judgement only gets trained on questions that are real.

The translation ladder Five rungs descending from a scientific question, to a measurable quantity, to a machine task, to an objective and metric, to a split and a control. A bracket marks that an AI agent can execute all five. Rung two, the measurable quantity, is highlighted as the one only a human can choose. AGENT RUNS ALL FIVE 01 The scientific question “Does this drug change how the cells look?” 02 A measurable quantity mean nuclear area per well, from DAPI 03 A machine task segment nuclei → measure → compare to control 04 An objective and a metric effect size + 95% CI, FDR across 84 genes 05 A split and a control hold out plate 4 · shuffle labels · buffer wells
Only a person who knows the biology can pick rung 02.
  1. 01

    The scientific question

    “Does this drug change how the cells look?”

  2. 02

    A measurable quantity

    Mean nuclear area per well, from DAPI. Only a person who knows the biology can pick this one.

  3. 03

    A machine task

    Segment nuclei → measure → compare to control.

  4. 04

    An objective and a metric

    Effect size + 95% CI, FDR across 84 genes.

  5. 05

    A split and a control

    Hold out plate 4 · shuffle labels · buffer wells.

An AI agent can run all five. It cannot tell you that rung 02 was the wrong quantity to measure.

What we actually run

Frontier AI agents, pointed at your problem

The whole course is about one thing: getting real scientific work out of an AI agent reliably. Not prompting tricks — specification, direction and verification.

A real coding agent

Not a chatbot in a browser tab. An agent that reads your files, writes and runs code, plots, and iterates on its own output until the task is done.

A person in the loop

Every agent run is specified, steered and checked by a course participant. They sign the result — “the agent said so” is not a methods section.

Controls, not vibes

Held-out splits, shuffled labels, buffer wells. The agent is fast and fallible, so checking it is part of the work rather than an afterthought.

This is what the course teaches, and what your project is a live exercise in. You do not need to know anything about agents to take part.

The procedure

Two meetings. Three weeks. One answer.

Everything in blue is time you spend. Everything else happens without you.

How a volunteered project runs, in five steps Step 1, register the project, five minutes. Step 2, meeting one, one hour online, where the student interviews you. Step 3, they build a prototype in under a week. Step 4, meeting two, one hour online, where you say what is wrong. Steps three and four repeat for one or two short rounds. Step 5, hand-over: code, figures and a written report, yours to keep. Your total time: two 1-hour meetings — about 2 hours, spread over three weeks further rounds are optional — by agreement between you and the participant We match you with a participant. If nobody is free, our AI data scientist runs it — every project gets an answer.

Your total time: two 1-hour meetings — about 2 hours, over three weeks

  1. Sign up

    2 minutes

    Tell us the question. Attach a sample file now if you like.

  2. Meeting 1

    1 hour · online

    They interview you — and this is where you hand the data over.

  3. They build it

    2–3 days

    An AI agent does the work; a person checks every number.

  4. Meeting 2

    1 hour · online

    You say what is wrong. More rounds only if you both want them.

  5. Hand-over

    yours to keep

    A written report and an analysis app you can run yourself.

We match you with a participant. If nobody is free, our AI data scientist runs it — every project gets an answer.

Your part, precisely

Exactly how it goes

No preparation, no slides, no code. Five things, in order — and two of them take an hour.

  1. Sign up

    2 minutes, now

    Not a form — an assistant asks you a few short questions and writes the details down for you. You can attach a sample file there if one is handy, but it is optional and most people do not.

  2. We match you with someone

    nothing to do

    We pair your problem with a participant on the course and introduce you by email. If nobody is free for your project, our AI data scientist takes it on instead — either way, you get an answer.

  3. Meeting 1 — the interview

    1 hour · online

    They ask you what you want to find out, how the data was produced, what you have already tried, and what would make you distrust a result. This is also where you hand the data over — so come with it ready, or with a way to share it. Recorded for teaching only if you agree.

  4. They build it

    2–3 days

    An AI agent does the implementation; the participant directs it and checks every number. You hear nothing for a couple of days, which is normal.

  5. Meeting 2 — you correct it

    1 hour · online

    They show you what they built, and your job is to say what is wrong, what is missing and what to drop. Being told it is wrong is the expected outcome. Further rounds are optional — only if you and they both want them.

  6. Hand-over

    nothing to do

    A written report, plus a small analysis app or the code itself — something that runs on your data after they are gone. Yours to keep, use, ignore or build on.

You need no programming and no AI experience. You bring the biology and the judgement about whether an answer is believable — the half of this that cannot be automated.

Why bother

What you get out of it

A working prototype

Code, figures and a pipeline that still runs after they are gone — plus a short written report.

Someone asking hard questions

An outsider has to ask what you never say out loud. That is often worth as much as the analysis itself.

An answer either way

If no student takes your project, our AI data scientist runs it. Nobody who volunteers goes away empty-handed.

No cost, no obligation

Free, no claim on your data, results or authorship. Use what they build, or don't.

What makes a good project

  • A real, unanswered question — not a demo, not something already solved.
  • Data that exists now and that you may share. A de-identified or subsampled extract is completely fine.
  • Something a motivated participant can move in under three weeks part-time. It does not need to be big data.
  • Imaging, omics, screens, clinical tables, sequencing, structures, awkward spreadsheets — all welcome.
  • Data that cannot leave your machine in any form — though tell us anyway, we may be able to arrange around it.
  • A question you already know the answer to. Nothing for either side to learn.
Timeline from registration to hand-over Register by 25 September, matched in late September, first meeting in early October, prototype and second meeting during October, hand-over by 23 October. Friday 25 September deadline to register a project Late September we match you and introduce you by email Early October meeting 1 — the interview · 1 hour Mid October meeting 2 — the prototype · 1 hour By 23 October hand-over — code, figures, report
  1. Friday 25 September

    Deadline to sign up a project.

  2. Late September

    We match you and introduce you by email.

  3. Early October

    Meeting 1 — the interview · 1 hour.

  4. Mid October

    Meeting 2 — the prototype · 1 hour.

  5. By 23 October

    Hand-over — report, figures, app.

Two routes, one promise A registered project goes either to a student or, if none is free, to the course's AI data scientist. Both routes converge on the same hand-over, checked by the teaching team. Your project REGISTERED A student the normal route AI data scientist if no student is free Your answer BOTH ROUTES CHECKED BY THE TEACHING TEAM
  • RegisteredYour project
  • The normal routeA course participant A master's student, PhD student, postdoc or researcher on the course.
  • If nobody is freeOur AI data scientist It interviews you in the same meeting and does the same work.
  • Either way Your answer Both routes checked by the teaching team.

Our side of the bargain

Every project gets an answer

We would rather have more projects than students. If nobody takes yours, our AI data scientist runs it — it interviews you in the same meeting, does the same work, and a member of the teaching team checks everything before it reaches you.

2 × 1 hmeetings, online
~3 weeksstart to hand-over
0 krand no obligation

Register

Hand us the question you never had time to answer

Signing up takes about two minutes and gives you a private project page — upload your sample file there, watch the work move through its stages, and raise anything that comes up while it runs.

Not sure yet? The assistant answers questions too — ask it there before you commit to anything.

Deadline: Friday 25 September 2026 · contact the course team