Applied ML Engineer: From Research to Production

Jobgether SRL; Netherlands; On-site; EUR 70,000 - 110,000; Full time

Benefits offered by this job

Location: Romania

Open-weight models exposure

Production-grade tooling

Job summary

Jobgether is seeking an Applied ML Engineer based in the Netherlands to bridge ML research and production. You will evaluate methods, design experiments, and build robust verification tooling across model weights, APIs, and inference infra while collaborating across backend, frontend, and product teams.

The role emphasizes hands-on work with PyTorch, open-weight models, and scalable evaluation infrastructure, with a strong focus on reproducibility, instrumentation, and clear reporting.

Qualifications

• Strong Python engineering skills with hands-on experience using PyTorch and Hugging Face Transformers.
• Solid understanding of ML evaluation including dataset design, baselines, metrics, calibration, and reproducibility.
• Ability to read ML research papers and implement methods from first principles.
• Professional software engineering experience beyond notebooks including APIs, asynchronous jobs, databases, logging, testing, deployment, and documentation.
• Familiarity with open-weight models and practical understanding of LLM inference systems.
• Ability to work across backend and frontend boundaries with sufficient React/TypeScript knowledge to help make experiments and results understandable to users.
• Ownership and initiative, with ability to drive projects independently.
• Experience with evaluation and inference infrastructure such as DSPy, LiteLLM, Temporal, Ray, vLLM, PostgreSQL/pgvector, or comparable technologies.
• Experience running and serving open-weight models on GPUs, focusing on latency, throughput, memory, precision, and cost.
• Experience designing adversarial evaluations or testing systems against evasion attempts.

Responsibilities

• Reproduce and evaluate ML research methods using open-weight models and APIs.
• Design evaluation datasets, probes, scoring approaches, baselines, calibration tests, and experiment harnesses.
• Work with model weights, logits, hidden states, and inference infrastructure as needed.
• Build and extend evaluation infrastructure including runners, judges, persistence, orchestration, reporting, and reproducibility.
• Translate research workflows into product experiences with experiment config, execution, traces, comparisons, reports, and review flows.
• Investigate verification methods under fine-tuning, merging, quantization, distillation, safety removal, or evasion.
• Create controlled experiments distinguishing signal from artifacts and confounding factors.
• Produce clear technical reports separating evidence from interpretation.
• Deliver production-quality systems with APIs, asynchronous jobs, databases, observability, testing, deployment, and documentation.
• Collaborate across research, experimentation, engineering, and product as priorities evolve.
• Within six months, reproduce and document at least one model-provenance or verification method and its capabilities and limitations.
• Build a repeatable model-verification runner with versioned inputs, artifacts, metrics, and reports accessible via the product interface.
• Run controlled experiments across base, fine-tuned, merged, quantized, and distilled models to improve understanding of verification methods.

Skills

PyTorch

Hugging Face Transformers

API Development

Asynchronous Jobs

Databases

Logging

Testing

Documentation

React/TypeScript

Next.js

Temporal

Ray

vLLM

pgvector

PostgreSQL

Tools

PyTorch

Hugging Face Transformers

React/TypeScript

Next.js

Temporal

Ray

vLLM

pgvector

PostgreSQL

DSPy

LiteLLM

Job description

Jobgether is seeking an Applied ML Engineer based in the Netherlands to bridge ML research and production. You will evaluate methods, design experiments, and build robust verification tooling across model weights, APIs, and inference infra while collaborating across backend, frontend, and product teams.

The role emphasizes hands-on work with PyTorch, open-weight models, and scalable evaluation infrastructure, with a strong focus on reproducibility, instrumentation, and clear reporting.