Senior ML Engineer, Post-Training & Production AI

Lila Sciences, Inc.; San Francisco (CA); On-site; USD 252,000 - 336,000; Full time

Benefits offered by this job

Medical, dental, and vision coverage

Employer-paid life and disability保险

Generous paid time off and holidays

Parental leave

Educational assistance program

Commuter benefits

Company lunch program

Job summary

Lila Sciences is hiring a Principal Machine Learning Engineer, Applied AI to own model post-training, evaluation, and production ML systems. You’ll shape how Lila’s models adapt to scientific customer needs and turn frontier capabilities into reliable workflows used in real contexts.

You’ll collaborate with AI researchers and software teams, build end-to-end evaluation loops, mentor engineers, and drive how model behavior is tuned through post-training and tooling across the product.

Qualifications

• Hands-on post-training experience with SFT and RL methods and designing evaluation systems.
• Strong software engineering skills in Python and PyTorch.
• Experience leading cross-functional research and engineering teams.

Responsibilities

• Close the gap between model capabilities and customer workflows.
• Lead post-training efforts using SFT and RL to align models with requirements.
• Build evaluation loops measuring quality, reliability, and customer fit.
• Collaborate with researchers and software teams to deploy production ML systems.
• Mentor engineers and share reusable patterns for model adaptation and deployment.

Skills

Post-training ML expertise

Python

PyTorch

RL methods (DPO, PPO, GRPO)

Model evaluation

Production ML systems

Mentoring

Job description

Lila Sciences is hiring a Principal Machine Learning Engineer, Applied AI to own model post-training, evaluation, and production ML systems. You’ll shape how Lila’s models adapt to scientific customer needs and turn frontier capabilities into reliable workflows used in real contexts.

You’ll collaborate with AI researchers and software teams, build end-to-end evaluation loops, mentor engineers, and drive how model behavior is tuned through post-training and tooling across the product.