ML Infrastructure Engineer — Data & Training

Arena Physica; New York (NY); On-site; USD 150,000 - 230,000; Full time

Benefits offered by this job

Premium medical, vision, dental

401(k) plan

Unlimited PTO

Relocation support

Lunch from local restaurants

Job summary

Arena Physica is building a team to advance electromagnetic superintelligence. You will join the Platform team as a Machine Learning Engineer focused on Data & Training Infrastructure, shaping how simulation data, measurements, and expert workflows feed scalable training data for the first electromagnetic foundation model.

Work closely with applied researchers, electrical engineers, and product engineers to orchestrate solver farms, data corpora, and fast, reproducible model development in a

Qualifications

• 5+ years of software engineering or ML infrastructure experience at a venture-backed startup or high-performing research org.

Responsibilities

• Scale electromagnetic training data generation infrastructure.
• Design dataset APIs, schemas, lineage systems, and quality gates.
• Own high-throughput training pipelines for large models.
• Collaborate with researchers to translate model needs into data programs.
• Make ML experiments reproducible with dataset snapshots and traceability.
• Integrate EM solvers and measurements into production-grade pipelines.

Skills

Python

ML infrastructure

Distributed systems

Data processing

Experimentation & reproducibility

Communication

Ownership mindset

Ambiguity handling

Collaboration

Hardware/EM knowledge (preferred)

Tools

PyTorch

JAX

Slurm

Ray

Kubernetes

AWS Batch

Elastic Fabric Adapter

Job description

Arena Physica is building a team to advance electromagnetic superintelligence. You will join the Platform team as a Machine Learning Engineer focused on Data & Training Infrastructure, shaping how simulation data, measurements, and expert workflows feed scalable training data for the first electromagnetic foundation model.

Work closely with applied researchers, electrical engineers, and product engineers to orchestrate solver farms, data corpora, and fast, reproducible model development in a