SlipstreamJobs tracks this role from the company's public career site. Apply directly on the employer's site.
Salary: USD 116,000 - 170,000 / annual
Lila Sciences is seeking Machine Learning Engineers to join their Applied AI organization, which sits at the intersection of AI research, model engineering, and product deployment. The role focuses on bridging research and engineering to turn frontier AI model capabilities into reliable, production-ready workflows for customer-specific scientific applications.
You will work on closing the last-mile gap between Lila's AI capabilities and real-world scientific workflows. Key responsibilities include post-training models using techniques like SFT, DPO, PPO, and GRPO to align behavior with customer requirements; building evaluation loops to measure model quality and reliability; designing experiments to improve performance across applied use cases; and debugging model failures using traces, evaluations, and customer feedback.
The role requires strong collaboration across teams: partnering with AI researchers to translate improvements into usable capabilities, working with software engineers to integrate models into end-to-end product workflows, and communicating customer needs back into technical model improvements.
You'll need solid software engineering fundamentals in Python and modern ML frameworks (PyTorch, JAX, TensorFlow), experience designing experiments and evaluation metrics, and the ability to debug model behavior using data and qualitative feedback. Familiarity with large language models, multimodal models, or agentic AI systems is expected. Bonus experience includes adapting models for production workflows, building evaluation harnesses and monitoring dashboards, working with retrieval-augmented generation or tool use, RL post-training techniques, MoE architectures, and translating product/customer needs into ML improvements.
This is an early-stage opportunity with autonomy and compute resources to tackle frontier science problems. The team values clear communication and the ability to move promising capabilities from research into production-quality systems.
