Senior Architect, Model Performance and Post-Training

What you'll do

• Build frameworks to evaluate coding models across code generation, debugging, code completion, repository reasoning, and multi-step software engineering.
• Develop and optimize post-training pipelines, including supervised fine-tuning, preference optimization, reinforcement learning, and reward modeling.
• Analyze model behavior and performance, design targeted training improvements, and build reproducible training and evaluation infrastructure.

What they ask for

• Bachelor's degree (or equivalent experience) in Computer Science or AI.
• Excellent Python programming skills and strong understanding of LLMs, transformer architectures, and generative-model training and evaluation.
• Foundational understanding of reinforcement learning algorithms and implementation, including MDPs, policies, and reward design.
• Deep understanding of software engineering fundamentals, code generation, automated testing, and program correctness; track record leading model development life cycles.

Nice to have

• Experience training or post-training state-of-the-art coding models or software engineering agent models.
• Experience with execution-based or verifiable rewards, reward hacking mitigation, automated code evaluation, or coding benchmarks such as SWE-bench.

Pay

224,000 USD - 356,500 USD (base salary); 272,000 USD - 431,250 USD (base salary)

Tools

Python, SWE-bench

AI focus

Performance modeling, evaluation, and post-training of LLMs for code generation, software engineering, and agentic coding; includes SFT, preference optimization, RL, reward modeling, and benchmark analysis.