AI Researcher (Post Training)

TL;DR

AI Researcher (Post Training): Developing and optimizing the post-training pipeline for a natural-language software creation platform with an accent on RFT/RLVR, preference optimization, and code generation. Focus on translating cutting-edge research into production-ready training recipes and building scalable evaluation systems.

Location: Stockholm, Sweden

Company

hirify.global enables millions of users to transform raw ideas into real software products using plain language.

What you will do

• Own the full post-training lifecycle, from data curation and training runs through evaluation and deployment.
• Adapt reinforcement learning, preference optimization, and SFT to improve code generation, reasoning, and agentic reliability.
• Build evaluation and experimentation infrastructure to measure helpfulness, safety, latency, and reliability.
• Develop and operate production systems for large-scale training, including GPU orchestration and data pipelines.
• Collaborate with agent, product, and infrastructure engineers to translate model gains into user-facing improvements.
• Investigate and resolve end-to-end failures in training recipes, data, or serving regressions.

Requirements

• Hands-on experience running post-training jobs (RFT/RLVR, preference optimization) on LLMs.
• Ability to write reliable production-grade code.
• Proficiency in PyTorch or JAX and experience with distributed training and GPU clusters.
• Strong understanding of the mathematics behind reward modeling, alignment, and preference optimization.
• Experience building evaluation systems that capture real-world quality rather than just benchmarks.
• English: Required (company language)

Nice to have

• Experience with code generation or agentic use cases.
• History of owning the full loop from data curation to production monitoring.
• Ability to rapidly prototype research papers into running code.
• Experience with speculative decoding or other model efficiency techniques.
• Contributions to the open-source ML ecosystem or research publications.

Culture & Benefits

• Talent-dense team with a culture of extreme ownership and high velocity.
• Low-ego collaboration environment.
• Fast-paced atmosphere focused on shipping impact to users quickly.