Researcher, Post Training

Responsibilities

• Own the full lifecycle of Lovable's post-training pipeline, from data curation and training runs through evaluation and deployment
• Apply and adapt reinforcement learning, preference optimization, and supervised fine-tuning methods for code generation and agent workloads
• Build evaluation and experimentation infrastructure to measure helpfulness, safety, latency, and reliability
• Develop and operate production systems for large-scale training jobs, including GPU orchestration and data pipelines
• Collaborate with agent, product, and infrastructure engineers to translate model gains into product improvements
• Investigate and resolve failures across training recipes, data issues, or serving regressions

Requirements

• Experience running post-training jobs on LLMs (RFT/RLVR, preference optimization, etc.)
• Ability to write solid production-grade code
• Fluency in PyTorch or JAX
• Experience with distributed training setups and GPU clusters
• Deep understanding of preference optimization, reward modeling, and alignment techniques
• Experience building evaluation systems that capture real-world quality
• Ability to trace model quality regressions from user symptoms back to training or inference

Preferred Qualifications

• Experience with code generation or agentic use cases
• Experience deploying post-trained models to real users at scale
• Experience owning the full loop: data curation, training, evaluation, deployment, and monitoring
• Experience with speculative decoding or model efficiency techniques
• Strong views on evaluation methodology and building predictive evals
• Contributions to the open-source ML ecosystem

About the Company

Lovable lets over 2 million people build software using plain language. We are a small, talent-dense team building a generation-defining company from Stockholm, focused on extreme ownership and high velocity.