The Senior AI/ML Engineer will lead the design, training, and deployment of large-scale machine learning models at Absentia Labs, working at the intersection of model architecture, training systems, and production infrastructure. This role involves significant ownership over technical direction and requires deep expertise in scaling models across complex scientific domains.
What You'll Do
- Design, train, and evaluate large-scale models, including Large Language Models (LLMs), diffusion models, and Graph Neural Networks (GNNs)
- Own end-to-end training pipelines, from dataset interfaces and batching strategies to distributed training and checkpointing
- Make principled decisions about model architecture, objective functions, optimization strategies, and scaling laws
- Build and optimize distributed training systems (data parallelism, model parallelism, sharding, mixed precision)
- Collaborate closely with data engineers to define ML-ready datasets and streaming interfaces
- Translate ambiguous scientific or product requirements into robust ML solutions
- Drive model evaluation, ablation, and iteration with a focus on generalization, stability, and reproducibility
- Contribute to architectural decisions around model serving, inference efficiency, and lifecycle management
- Provide technical leadership through design reviews, mentorship, and cross-team collaboration
What We're Looking For
- 5+ years of industry experience in machine learning or applied AI roles
- Demonstrated experience training large-scale models in production settings, not just prototypes
- Hands-on expertise with LLMs, diffusion models, and/or GNNs
- Strong proficiency in PyTorch (or equivalent deep learning frameworks)
- Deep understanding of distributed training, including parallelism strategies and performance optimization
- Experience working with large datasets and high-throughput data pipelines
- Strong software engineering fundamentals: clean code, testing, reproducibility, and debugging at scale
- Ability to clearly communicate technical trade-offs to both technical and non-technical stakeholders
Nice to Have
- Experience with reinforcement learning, fine-tuning, or preference-based optimization (e.g., RLHF)
- Familiarity with model compression, distillation, or inference optimization
- Experience deploying models in production inference systems
- Exposure to multimodal learning or foundation models
- Prior work in startups or fast-moving R&D environments
- Contributions to open-source ML frameworks or research codebases
Technical Stack
- PyTorch
- Large Language Models (LLMs)
- Diffusion Models
- Graph Neural Networks (GNNs)
- Distributed Training
- Data Parallelism
- Model Parallelism
- Sharding
- Mixed Precision Training
Benefits & Compensation
- Competitive compensation
- Meaningful equity participation
- Opportunity to work on foundation-level ML systems applied to real scientific problems
- Ownership over model design and training strategy, not just implementation
- Close collaboration with data, infrastructure, and scientific teams
- High autonomy
- Low bureaucracy
- Culture that values technical depth
- Flexible remote or hybrid work arrangements
Work Mode
Flexible remote or hybrid work arrangements
Absentia Labs is an equal opportunity employer. We believe diverse teams build better systems and stronger science, and we encourage applicants from all backgrounds to apply.
