New comment by lyluo in "Ask HN: Who is hiring? (November 2026)"

Status
Open
Remote policy
Hybrid
Employment type
Not stated
Salary
Not stated
Tech
pythonhybrid
Source
hn-whoishiring
First observed
2026-11-13 09:57 UTC
Last seen
2026-11-13 09:57 UTC
Source claims posted
2026-11-13 09:04 UTC
Consecutive misses
0 of 10

What the posting says

CERN | Machine Learning Engineer | Geneva, Switzerland | HYBRID | Full-time, 3-year contract | From CHF 9,236/month, tax-free

Not a physics job: this role is in CERN's healthcare AI team. CAFEIN, CERN's federated AI platform, trains models on medical data held by hospitals and research institutes across Europe without that data ever leaving them. It powers EU projects such as UMBRELLA and AIM-SAFE.

It's hands-on ML engineering, from experiment to production:

- designing and running ML experiments in Python (PyTorch, TensorFlow, JAX, scikit-learn)

- deploying models on distributed infrastructure shared by several institutions

- privacy and security: differential privacy, encryption, access control

- tools for data ingestion and data quality, including medical images

- benchmarking state-of-the-art methods for accuracy, efficiency and scalability

Looking for an ML engineer who has taken models to production and is comfortable with Python, PyTorch/TensorFlow/JAX, Git/GitLab and the security side of distributed systems. Experience with federated or distributed learning, or with healthcare data, is a big plus. Master's in ML/CS/Data Science or equivalent experience. English required, plus a commitment to learning French.

Eligibility: open to nationals of CERN Member and Associate Member States.

Tax-free salary from CHF 9,236/month, 30 days of leave plus a two-week annual closure, relocation package, health insurance and pension fund. The contract is 3 years, extendable up to 8, with the possibility of applying for an indefinite position.

Apply by 12 December 2026: https://jobs.smartrecruiters.com/CERN/744000161227765-machin...

Quality

Completeness: 50%

Not enough history yet to judge honesty signals.

Timeline

  1. *
    #1897367 2026-11-13 09:57 UTC
    Published