Senior ML Researcher (AI)

Location: European Union. Full remote is available; hybrid work is available if based in the Netherlands or Serbia.

Company

hirify.global AI builds data and machine-learning products that help leading AI companies train, evaluate, and improve generative AI models.

What you will do

• Own end-to-end fine-tuning pipelines, including data preparation, SFT/LoRA training, model distillation, evaluation, and serving handoff.
• Extend the post-training stack with reinforcement learning, GRPO-style methods, reward modeling, and LLM-judge-based rewards.
• Build evaluation harnesses calibrated against human labels, including golden datasets and regression evaluations.
• Improve the platform’s guiding agent through prompt and tool design, evaluation-driven iteration, and stress testing.
• Run experiments for client projects and convert successful approaches into repeatable platform capabilities.
• Collaborate with platform engineers on SDK and API interfaces for training and evaluation workflows.

Requirements

• 4+ years of experience in ML engineering or applied research, including 1–2 years working hands-on with LLMs.
• Practical experience fine-tuning open-weight models with LoRA or full fine-tuning, including data curation.
• Strong understanding of LLM evaluation and LLM-as-judge calibration against human judgments.
• Strong Python engineering skills and experience with PyTorch, the Hugging Face ecosystem, and vLLM or similar tools.
• Product mindset and ability to work effectively with ambiguity and changing priorities.
• English fluency at B2 level or above; candidates must be based in the European Union.

Nice to have

• Experience with RLHF, RLAIF, GRPO, PPO-style post-training, or reward modeling.
• Experience with DSPy, GEPA, prompt compression, distillation, or quantization.
• Experience building agentic systems or shipping ML features in self-serve products.

Culture & Benefits

• International, remote-first environment with a globally distributed team.
• Full remote or hybrid work model, with hybrid availability in the Netherlands and Serbia.
• Competitive compensation package with base salary, bonus, and ESOP.
• Paid PTO, location-dependent benefits, IT setup, and home office allowances.