Train the models. Then prove them.
We build the models that decide which products each shopper sees, then test them on live traffic. We work from Cape Town.
Open roles
- Machine Learning ResearcherResearch and train the multimodal models behind Solenya's product discovery, and work out how to prove they're better.Full-time · Cape Town, South Africa
- Machine Learning EngineerTake our models from research to production, and keep them fast and reliable at millions of requests a day.Full-time · Cape Town, South Africa
What the work is like.
- Your work ships.Research ends as code in production, not only as a paper.
- The eval comes first.Whether the eval is right matters more than the number it gives.How we test changes →
- A small team of generalists.Model serving, data pipelines, the merchant API and the infrastructure under them.
- You own what you build.You work directly with the leadership team.
What you'd work in.
- ML
- PyTorchHugging FaceJAX/FlaxRay (Train, Serve, Data)SGLangMLflow
- Backend
- FastAPIGraphQL (Strawberry)PrefectKafka
- Data
- DVCBigQuerydbtPolarsMilvusPostgresClickHouse
- Infra
- GKE with GPUsHelmSkaffoldTerraformNixuv
Who you'd work with.
Meet the team →Don't see your role? Write to us anyway.
Send your GitHub, or whatever shows your work best. The address is in the secret.
Or skip the decoding →apiVersion: v1
kind: Secret
metadata:
name: careers
type: Opaque
data:
email-address: Y2FyZWVyc0Bzb2xlbnlhLmFp