Member of Technical Staff - Post Training, Applied

ABOUT LIQUID AI

Spun out of MIT CSAIL, we build general-purpose AI systems that run efficiently across deployment targets, from data center accelerators to on-device hardware, ensuring low latency, minimal memory usage, privacy, and reliability. We partner with enterprises across consumer electronics, automotive, life sciences, and financial services. We are scaling rapidly and need exceptional people to help us get there.

THE OPPORTUNITY

You will own applied post-training for text workloads, adapting Liquid Foundation Models for some of the world’s largest enterprise customers. Working directly with customers and our internal technical teams, you will translate business needs into models that perform reliably on real tasks, with ownership from scoping through adaptation, evaluation, and delivery. Between engagements, you will turn what you learn into reusable tooling and workflows that make future delivery faster and more effective.

WHAT WE'RE LOOKING FOR

We need someone who:

- Takes ownership: Clarifies ambiguous requirements, surfaces risks early, and follows through until the customer’s problem is solved.

- Lets evidence guide decisions: Tests assumptions, investigates failure cases, and chooses approaches based on measurable improvements in model quality and customer outcomes.

- Communicates clearly: Explains technical trade-offs, keeps customers and internal teams aligned, and pushes back when expectations or proposed approaches need to change.

THE WORK

- Own enterprise post-training engagements for workloads, translating customer requirements into technical specifications, delivery plans, and evaluation criteria.

- Build data generation, curation, filtering, and quality assessment workflows tailored to customer tasks.

- Select and execute appropriate post-training methods, including supervised fine-tuning, instruction tuning, and preference alignment approaches such as RLHF and DPO.

- Design task-specific evaluations, analyze model failures, and use the results to guide improvements in data and training.

- Deliver adapted models with clear evidence of performance, documented limitations, and guidance for customer use.

- Build reusable tooling and workflows that accelerate future engagements and feed practical insights back into core model development.

EXPERIENCE & CAPABILITIES

What you'll need:

- Particularly strong judgment in data quality and evaluation design, supported by hands-on experience generating, curating, and evaluating data for LLM post-training.

- Experience training or fine-tuning language models using methods such as SFT, instruction tuning, RLHF, DPO, or similar preference alignment approaches.

- The ability to reason across data, training, alignment, and evaluation as a connected system, with practical experience in chat model alignment, instruction tuning, or text data curation at scale.

- Proficiency with PyTorch, Hugging Face, and modern language model architectures.

Helpful, not required:

- Experience delivering applied ML projects to external customers with measurable outcomes.

- Familiarity with inference optimization frameworks such as vLLM, SGLang, or TensorRT.

- Experience building reusable ML tooling or evaluation infrastructure.

WHAT SUCCESS LOOKS LIKE (YEAR ONE)

- You have independently delivered enterprise text post-training projects that meet agreed evaluation criteria, with minimal oversight.

- Customers trust your technical judgment and delivery, with a clear understanding of model performance, limitations, and trade-offs.

- Your reusable tooling and workflows are being used in subsequent engagements, reducing repeated work and accelerating delivery.

WHAT WE OFFER

- Applied ML with direct impact: Generate data, fine-tune models, and ship solutions for enterprise customers. Your work feeds directly back into our core model development.

- Compensation & Equity: Competitive base salary plus equity, so you share in what we build. Final offers reflect skills, experience, and location.

- Time Off: Unlimited PTO, plus company-wide Refill Days, long weekends when the whole company rests together.

- Health & Wellbeing: Comprehensive health coverage, with details varying by location.

- Paid Leave: Paid leave for pregnancy, parenting, medical needs, and caring for family.