Metly is a pre-seed startup (team of 5) building strategic intelligence for pharmaceutical companies. We're aiming to replace consultants.
Pharma companies hold critical knowledge across disconnected systems—regulatory filings, clinical pipelines, competitive intelligence—none of it linked to real market activity. We're building the layer that connects it all and makes it actionable.
You'll be working on the hard parts of getting LLMs to construct knowledge graphs users can actually trust: entity extraction from dense regulatory documents, grounding outputs with verifiable citations, and turning unstructured data into clear recommendations.
Stack: TypeScript/Python, React on TanStack, PostgreSQL, Kubernetes on GCP.
We have traction and are moving fast in 2026. Competitive salary and meaningful early-stage equity.
I don't get the hype for a specific agent here really - most cloud sdks _+ pulumi support languages like typescript and python for IaC. No YAML. I think terrform was thinking about this the last time I was working with them. That gives you all the benefits of software agents we already have available. I think for most those will be enough.
Right — and that’s exactly what people said about shell scripts before IaC, and about IaC before SDK-based infra. Each wave solves a pain point, then new complexity emerges. Agents are starting to address the next layer of that complexity: intent, adaptation, governance, and collaboration.
Excellent book on mathematical thinking in the true sense - what needs to happen in the mind's eye to really grapple with abstract mathematics. Definitely a eye (mind?) opener for someone who has some graduate level math education but couldn't gel with the crazier stuff.
Unspun is cool! I hadn’t heard of it before. They provide 3D weaving on demand, according to the site, which is the “other” clothes manufacturing technique — knitting is how, e.g. a sweater or more finely made t-shirts would be made, and the structure of the ties and loops in knitting provides that ‘stretch’. Weaving on the other hand shouldn’t stretch much, and would be used for, say, your curtains, jeans, or a dress that needs to fit exact sizes.
There are Japanese (and probably other) 3D in the round knitting providers, but I hadn’t heard of weaving done on demand and to spec — that’s pretty amazing.
I would recommend "The Things We Make" for an outlook into engineering mindsets through history. A good reminder that a lot of useful engineering comes before the theory can fully explain it :)
I'm free of this pain currently but I could see this saving a huge time pain in the product -> engineering handover step. Very cleanly done. Honestly great just for making sure engineering designs have covered all paths.
Metly is a pre-seed startup (team of 5) building strategic intelligence for pharmaceutical companies. We're aiming to replace consultants. Pharma companies hold critical knowledge across disconnected systems—regulatory filings, clinical pipelines, competitive intelligence—none of it linked to real market activity. We're building the layer that connects it all and makes it actionable.
You'll be working on the hard parts of getting LLMs to construct knowledge graphs users can actually trust: entity extraction from dense regulatory documents, grounding outputs with verifiable citations, and turning unstructured data into clear recommendations.
Stack: TypeScript/Python, React on TanStack, PostgreSQL, Kubernetes on GCP. We have traction and are moving fast in 2026. Competitive salary and meaningful early-stage equity.
graeme(at)meetmetly.com
reply