Job Description
We are looking for a qualified Prompt Engineer and LLM Integration specialist with experience in optimizing interactions with Large Language Models (LLMs), including ChatGPT, through Python. The idóneo candidate will have a deep understanding of OpenAI’s APIs, experience in token management, and the ability to work with multiple LLMs and prompts to achieve the best results. A high level of experimentation and analysis is required for this position. This is a 12‑month opportunity.

Responsibilities

Token Management: Optimize token usage in interactions with LLMs, taking into account context limitations and associated costs.

API Integration: Develop and maintain Python code that efficiently interacts with OpenAI’s APIs to access different LLMs.

LLM Experimentation: Evaluate and compare the performance of different LLMs and prompt configurations to identify the best solutions for each use case.

Multiple Prompt Management: Design strategies for managing and executing multiple prompts in parallel or sequentially, including combining results from different LLMs.





Prompt Engineering: Design, refine, and optimize prompts for different LLMs (including GPT‑3.5‑turbo, GPT‑4, and potentially other models) with the goal of obtaining accurate, relevant, and coherent responses for a variety of tasks.

Results Analysis: Analyze the responses from LLMs, identifying patterns and areas for improvement in prompts and model configurations.

Documentation: Document experiments, results, and best practices related to prompt engineering and LLM integration.

Qualifications

Python: Proficiency in Python (version 3.7 or higher) and experience with libraries such as OpenAI, requests, pandas, and numpy.

Experience with OpenAI APIs: Familiarity with using OpenAI APIs to interact with ChatGPT and other LLMs.

Knowledge of LLMs: Solid understanding of how large language models work, including transformers, attention mechanisms, and encoding.

Prompt Engineering: Ability to de

📌 Process Engineer (Madrid)
🏢 Chubb
📍 Madrid