29 days ago
Palo Alto, CA, USAMid Level
Responsibilities
- Lead instruction tuning, preference tuning, and model alignment for production text workloads.
- Own customer post-training projects from requirements and data generation through training, evaluation, and delivery.
- Customize open-source models and transition them from post-training into production serving.
- Improve inference-time efficiency, reliability, robustness, latency, concurrency, and scalability for deployed models.
- Build reusable ML tooling, workflows, and evaluation infrastructure.
- Provide technical mentorship and guidance to the engineering team.
Requirements
- At least 2 years of experience building and deploying machine-learning-based services in production.
- Hands-on experience with data generation and evaluation for LLM post-training.
- Experience training or fine-tuning models using SFT, instruction tuning, RLHF, DPO, or similar preference-alignment methods.
- Strong understanding of text-data quality and evaluation design, including chat-model alignment, instruction tuning, or large-scale text-data curation.
- Proficiency with the open-source ML ecosystem, including Hugging Face and PyTorch, and modern model architectures.
- Preferred experience optimizing inference for on-premise deployments, improving deployed-system performance, serving low-precision models, using Multi-LoRA, or distributing models across GPU nodes.
- Preferred experience with large-scale production systems, open-source ML projects, Kubernetes ML workloads, external applied-ML delivery, inference optimization frameworks, or reusable ML tooling.
Tech Stack
Categories
About Sanas.ai
Sanas builds real-time speech AI that translates accents and enhances clarity on live conversations, helping multilingual agents be more easily understood. It sells the technology as SaaS via APIs, call-center integrations, and an agent desktop app to contact centers, BPOs, and enterprise support and sales teams. The company is privately held and headquartered in Palo Alto, California.