Autonomous Decisioning ML Researcher

Soledad Data Solutions; California (MO); On-site; USD 120,000 - 180,000; Full time

Job summary

Mylitm is seeking a Data & ML Researcher who sits at the intersection of research and engineering to advance autonomous decisioning in high-stakes domains. You will design rigorous experiments, translate results into production prototypes, and collaborate with MLOps to deploy.

You will work with modern ML techniques including LLMs, reinforcement learning, and structured prediction, shaping the research roadmap for one or more product surfaces and publishing internal notes when appropriate.

Qualifications

• PhD or equivalent in ML, statistics, or a related quantitative field.
• Strong publication record or evidence of shipped applied-research work.
• Fluent in PyTorch or JAX; comfortable with distributed training.
• Strong software engineering fundamentals — you can take a notebook to a service.

Responsibilities

• Design and run rigorous experiments on candidate model architectures and decisioning algorithms.
• Translate findings into production-grade prototypes and partner with MLOps to deploy.
• Read, summarize, and reproduce relevant academic and industry research.
• Build evaluation harnesses, eval datasets, and metrics that reflect real customer outcomes.
• Publish internal research notes and contribute to external publications when appropriate.

Skills

Strong publication record

Technical writing

De-risking ideas with experiments

Education

PhD or equivalent in ML/related field

Tools

PyTorch

JAX

Distributed training

Job description

Mylitm is seeking a Data & ML Researcher who sits at the intersection of research and engineering to advance autonomous decisioning in high-stakes domains. You will design rigorous experiments, translate results into production prototypes, and collaborate with MLOps to deploy.

You will work with modern ML techniques including LLMs, reinforcement learning, and structured prediction, shaping the research roadmap for one or more product surfaces and publishing internal notes when appropriate.