About the role

Senior ML Engineer advancing innovation in small language models for League's healthcare platform. Focused on research and engineering in model development and applied AI systems.

Responsibilities

Design and implement experiments across fine-tuning, distillation, and optimization of small language models (1–10B parameters)
Rapidly prototype and evaluate new approaches to model performance, efficiency, and reasoning quality
Leverage modern tooling and AI-assisted workflows to accelerate iteration cycles
Build applied systems that connect models, data pipelines, and evaluation frameworks
Focus on “wiring together” components across model training, evaluation, and deployment workflows
Collaborate with engineering teams to transition promising experiments into production environments
Contribute to training data design, including curation, labeling strategies, and synthetic data generation
Work with data partners to explore AI-driven insights and improvements to model performance
Define and run experiments to assess model performance across accuracy, reasoning, and safety dimensions
Contribute to building lightweight evaluation frameworks and benchmarking approaches
Actively leverage AI tools (e.g., Copilot, LLM-assisted coding, research copilots) to improve productivity and experimentation speed
Document and share workflows that improve how the team builds and evaluates models
Partner with Product, Platform Engineering, and AI Orchestration teams to integrate models into real-world use cases
Communicate complex technical concepts clearly to cross-functional stakeholders

5+ years of hands-on experience in applied ML/AI engineering, with a focus on language model development, fine-tuning, or NLP systems.
Proven track record shipping fine-tuned or distilled LLMs/SLMs (1–10B parameters) to production.
Deep expertise in PEFT techniques — LoRA, QLoRA, adapter tuning — and model quantization and distillation pipelines.
Hands-on experience with RLHF/RLAIF, reward modeling, or safety alignment workflows.
Strong background in data curation, labeling pipeline design, and synthetic data generation.
Proficiency with model training frameworks and tooling: NeMo, Hugging Face Transformers, Axolotl, or equivalent.
Experience with model serving stacks: vLLM, Triton, or similar; familiarity with inference optimization techniques.
Comfort operating on cloud infrastructure (GCP, Vertex AI, AWS) and with GPU resource management.
Solid understanding of healthcare data privacy and safety requirements: HIPAA, FHIR, clinical ontologies.
Demonstrated ability to define and own evaluation frameworks — not just build models, but know whether they're working.
Strong technical communication skills; able to present complex model decisions clearly to cross-functional and executive audiences.
Bachelor's or graduate degree in Computer Science, Machine Learning, or equivalent experience.