AI Engineer (NLP)

About the Client:

The client is a forward-thinking organization specializing in cutting-edge technologies and applied research. As part of their ongoing projects, they are looking for a highly skilled Technical Specialist to support their March work package focused on research, comparison, and demonstration of advanced Retrieval-Augmented Generation (RAG) techniques using state-of-the-art (SOTA) approaches combined with new innovations. The client’s projects are centered around processing large volumes of complex documents (e.g., legal texts, government regulations, court cases), making this an exciting opportunity for those passionate about applied machine learning and natural language processing (NLP).

Job Description:

The client seeks a hands-on Technical Specialist with expertise in applied Large Language Models (LLMs) and RAG techniques. The role involves researching and developing demonstrator code to compare naive RAG with a blend of SOTA and new methods, using existing source data. This role is critical to progressing the work package while the rest of the team focuses on other areas of the project.

You will be responsible for developing, optimizing, and testing various LLM-based approaches, particularly in the context of document processing, including laws, regulations, and case studies. The ideal candidate will have significant experience working with LLMs, RAG techniques, and open-source machine learning frameworks.

Responsibilities:

• Conduct research to compare and contrast naive RAG approaches with advanced SOTA techniques and new methodologies.

• Develop demonstrator code to showcase the comparison between RAG approaches.

• Leverage existing data sources (mainly PDFs, legal documents, regulations) for practical implementation.

• Apply NLP techniques and LLMs to process and organize large datasets.

• Construct and optimize Knowledge Graphs (NLP/LLM-assisted) to enhance RAG and other advanced techniques.

• Benchmark and optimize LLM models and RAG workflows.

• Provide technical leadership on the use of LLMs and RAG techniques.

• Collaborate with the team to integrate your findings into the broader project objectives.

Key Skills and Experience:

The ideal candidate should possess practical experience with the following:

• Serving Open Source Software (OSS) LLMs in a production environment.

• Experience with Supervised Fine-Tuning (SFT) and Data Preprocessing Optimization (DPO).

• Data preparation techniques for large-scale NLP applications.

• Constructing and optimizing Knowledge Graphs using NLP or LLM techniques.

• Hands-on experience with Graph RAG techniques.

• Experience in function calling and tool usage within LLM frameworks.

• Working with multi-modal LLMs, including text, audio input/output, and images.

• Proficiency in Python programming, particularly in the context of machine learning and NLP.

• Ability to benchmark LLM models and assess performance.

Desirable Skills and Experience:

• Familiarity with or practical experience in prompt engineering and optimization.

• Experience with Test-Time Preference Optimization (TPO) techniques.

• Knowledge of KV cache-stuffing and compression methods to enhance LLM performance.

• Understanding of LLM observability tools and techniques for debugging and monitoring.

• Experience with model distillation for optimization and deployment.

• Expertise in Personal Information Masking and Reconstruction methods.

• Knowledge of test-time autoscaling and batching for model deployment.

• Familiarity with version control systems (e.g., Git) and checkpointing for model management.

Start Date: March (exact start date to be confirmed)

Duration: Temporary/Contract, with the possibility of extension depending on project needs.

Why You Should Apply:

This role offers the opportunity to work at the forefront of LLM and RAG research, contributing to innovative projects with real-world impact. You’ll collaborate with a dynamic team and have the chance to push the boundaries of current LLM applications in complex document processing.

Commitment to Diversity, Equity, and Inclusion:

The client is committed to fostering a diverse, equitable, and inclusive work environment. They encourage candidates from all backgrounds to apply.