At Servient, we're building Generative AI powered LegalTech. We're looking for a talented and motivated Large Language Model Developer to join our dynamic team and fine-tune and serve cutting-edge LLMs to bring efficiency to legal tasks. You'll be instrumental in optimizing these models for fine tuning and scalable inference capabilities on GPUs.

Responsibilities:

• Design and implement NLP algorithms to work in conjunction with LLMs to analyze large legal documents, help legal professionals find evidence quickly and organize their evidence.
• Collaborate with legal professionals to understand their workflows and come up with creative solutions to bring efficiency to these tasks
• Develop custom training pipelines for utilizing the power of GPUs in LLM for inference.
• Define and monitor various evaluation metrics for the NLP algorithms and continuously iterate to improve them.
• Collaborate with NLP scientists, AI engineers, software developers to integrate LLMs into our products for legal tasks.
• Stay up-to-date on the latest advancements in LLM research and development, actively contribute to knowledge sharing within the team.

Qualifications:

The ideal candidate should have 1-4 years of experience with a

• Strong understanding of Natural Language Processing (NLP) concepts and techniques, including deep learning models for language generation.
• Experience with supervised and unsupervised learning methodologies as applied to Natural language documents
• Expertise in deep learning frameworks (e.g., TensorFlow, PyTorch) and scikit-learn
• Excellent software development skills with a focus on code quality, performance optimization, and maintainability.
• Strong communication and collaboration skills, with the ability to work effectively in a fast-paced, research-oriented environment.

Nice to have:

• Experience in distributed computing and scaling machine learning models for production environments.
• A keen understanding of NLP applications in your specific field of interest (e.g., healthcare, education, finance).