Flag job

Report

Machine Learning Engineer - Fine Tuning

Salary

$150k - $225k

Min Experience

3 years

Location

San Francisco, New York

JobType

full-time

About the job

Info This job is sourced from a job board

About the role

As a Machine Learning Engineer specializing in Fine-Tuning at Baseten, you'll create exceptional value for customers by leveraging our world-class infrastructure to fine-tune large language models and/or other modalities, working directly with customers to achieve their specific goals. You'll build scalable pipelines, implement parameter-efficient techniques, and ensure a seamless transition to inference. This customer-facing role requires both technical expertise in foundation model adaptation and the ability to translate customer needs into effective solutions. You'll also help shape our product roadmap by identifying common patterns in customer requirements and working with product teams to develop reusable components and features , reducing the need for custom services and streamlining the fine-tuning process for everyone. RESPONSIBILITIES: Design comprehensive fine-tuning strategies that translate customer requirements into effective technical approaches—finding the optimal combination of data preparation, training techniques, and evaluation methods to deliver solutions that precisely address customer needs Develop tools to enable non-ML experts to fine-tune models effectively Design and implement scalable fine-tuning pipelines for large language models and other AI modalities Work directly with customers to understand requirements and guide technical implementation Serve as the technical point of contact for customers throughout their fine-tuning journey Utilize state-of-the-art parameter-efficient fine-tuning methods (LoRA, QLoRA) Build systems for efficient data preparation, evaluation, and deployment of fine-tuned models Research and apply cutting-edge techniques in instruction tuning and model customization Create frameworks to evaluate fine-tuned model performance against base models Implement best-in-class distributed training techniques like FSDP and DDP across various hardware configurations REQUIREMENTS: Bachelor's degree in Computer Science, Engineering, or related field 3+ years of experience in ML engineering with focus on model training and fine-tuning Experience with advanced fine-tuning frameworks such as Axolotl, Unsloth, Transformers, TRL, PyTorch Lightning, or Torch Tune, enabling efficient model adaptation and optimization Hands-on experience fine-tuning or pre-training LLMs or other foundation models Excellent communication skills for explaining complex concepts to varied audiences NICE TO HAVE: Experience working with customers to deliver technical solutions Track record of delivering ML projects to enterprise customers Knowledge of distributed training systems and efficiency optimization techniques Experience with advanced alignment and adaptation techniques including RLHF, DPO, constitutional AI, prompt tuning, reinforcement learning with execution feedback, PPO, or other emerging alignment methods Knowledge of prompt engineering and domain adaptation methods Contributions to open-source fine-tuning projects or tools Experience building user-friendly interfaces for fine-tuning workflows Experience with cloud platforms (AWS, GCP, Azure) and containerization technologies

About the company

Join our dynamic team at Baseten, where we're revolutionizing AI deployment with cutting-edge inference infrastructure. Backed by premier investors such as IVP, Spark Capital, Greylock, and Conviction, we're trusted by leading enterprises and AI-driven innovators—including Descript, Bland.ai, Patreon, Writer, and Robust Intelligence—to deliver top-tier performance, security, and reliability for their production workloads. With our recent $75 million Series C funding, we're poised to accelerate our mission to make AI accessible across all products.

Skills

machine learning
fine-tuning
large language models
ai modalities
data preparation
training techniques
model evaluation
parameter-efficient techniques
distributed training
llm
model customization
prompt engineering
domain adaptation