HMND / Life science × software / Est. 2019

Science moves faster than the software built to serve it.

HMND is the translation layer between research and engineering. We work with pharma and biotech teams to find the right question — before the engineers build the wrong answer.

01Embed domain experts 02Fix the question first 03Stay as the translation layer 04Human in the lead

01The cost of the wrong question

$0 Spent building the wrong thing before anyone asked the scientists.

  1. 01

    Three engineers. One product manager. One business analyst. Two months.

  2. 02

    A research team was building a test-result analysis suite. We were brought in to improve the UX of their report builder.

  3. 03

    One week of user research. One finding: the scientists needed to print batch results and tape them to the bench, next to the physical sample.

  4. 04

    The fix was a print button.

  5. Two months and $38,000 to ask the wrong question. One week to ask the right one.

02How this starts

Every engagement starts the same way.

A mapping engagement, with the scope and the price agreed before we begin. We spend it working out what your business actually contains, where your systems disagree about it, and which question you are really trying to answer. You get back a written architecture: what to build, in what order, what it costs you if it is wrong, and what it earns if it is right.

It is yours either way. If you never work with us again, you still have it.

Duration
Fixed. Agreed before we begin.
Price
Fixed. Agreed before we begin.
What you get back
A written architecture: what to build, in what order, what it costs if it is wrong, what it earns if it is right. The right question, written down.

Sometimes the honest answer is that you should not build the thing at all. We will tell you that, and it will be the cheapest advice you get all year.

Start with a mapping engagement

03Thesis

Intuition, made legible.

The knowledge in a researcher’s head is the most valuable and most fragile asset your organization owns. We build software that captures it, structures it, and makes it actionable — without destroying what made it valuable in the first place.

Since 2019 HMND has operated at the intersection of advanced science and software: through a pandemic, through platform shifts, through the arrival of generative models. Always at the edge of what is shippable, never past it.

04Method

How we work.

Four commitments. Every engagement. No exceptions negotiated at the kickoff.

  1. 01

    Embed domain experts

    Pharmacologists and research scientists sit inside the engagement, not in an advisory box. We co-architect with your experts, never around them.

  2. 02

    Fix the question first

    The problem is rarely the engineers. It is the brief. We find the wrong question, reframe it, and hand engineering something that is actually buildable.

  3. 03

    Stay as the translation layer

    We do not deliver a deck and leave. We remain a standing bridge between what the science demands and what the software can ship.

  4. 04

    Human in the lead

    Every AI-assisted workflow we design keeps a person in the loop with the authority to overrule. Losing agency to automation is the failure you cannot fix afterwards.

Teams we’ve shipped with

  • Takeda
  • Bruker
  • Biogen
  • GSK
  • Strateos
  • CDD Vault
  • Quantori
  • Azul
  • Ensysce
  • Synthesized

06Open problems

Where we are already building.

Active lines of work. If your problem sits near one of these, the conversation is short.

  1. 06.1Semantic modelling of research knowledge
  2. 06.2Small-molecule design — generative models and retrosynthesis
  3. 06.3Pharmacovigilance signal detection
  4. 06.4Patient-facing care tools that clinicians actually adopt

07Vision

Field notes from the frontier.

A running analysis of where science, software, and AI converge in pharma and biotech. No press releases.

Read Vision

08Contact

Describe your last six months. We’ll tell you which question was wrong.

Response
Within two working days.
Hours
local · GMT+3