Prompt Engineering, AI Training, LLM Fine-Tuning, Data Annotator

About this Gig

My career is defined by the unique intersection of Artificial Intelligence, linguistics, and engineering, which aligns perfectly with the requirements to design and implement intelligent agents specializing in LLMs, RAG, and Workflow Automation. My hands-on experience in training AI models with a focus on linguistic nuance and technical accuracy has given me a deep, foundational understanding of what makes an AI agent truly effective. My work in AI is directly relevant to the core responsibilities of this role: * Expertise in LLMs and Natural Language: In my role as a Gen AI Trainer at Telus International, I have extensive experience designing complex, reasoning-based prompts and sophisticated VQA prompts using LaTeX. This work goes beyond simple data entry; it involves a deep understanding of how LLMs process information to ensure both scientific and linguistic accuracy. Similarly, at Appen, I generated multilingual prompts to enhance the performance of conversational AI, which has given me practical insight into crafting systems that genuinely understand and respond to user intent. * Practical Understanding of RAG Principles: My experience as a Search Engine Evaluator at Lionbridge provided me with a robust framework for assessing the relevance, quality, and authority of information—the very principles that underpin effective Retrieval-Augmented Generation (RAG). I spent years analyzing how well retrieved information satisfies a user's query, giving me an intuitive grasp of the challenges and success factors in building RAG systems that deliver trustworthy and contextually appropriate results. * A Mindset for Workflow Automation: While my background is rooted in AI training and data quality, the systems I've worked with—from translation management with SDL Trados and MemoQ to large-scale data annotation—are all complex workflows. I excel at maintaining consistency, managing linguistic assets, and collaborating with QA teams to ensure quality across massive datasets. My engineering background further reinforces this systematic, process-oriented approach, which I am keen to apply to building and automating agentic workflows.

Requirements

To ensure high-quality output from the start, I require clear quality guidelines and success metrics, similar to the stringent standards I followed at Telus to achieve a 95% acceptance rate. I also need precise technical specifications, such as requirements for LaTeX formatting in STEM projects, and for linguistic tasks, access to existing Translation Memories and glossaries is essential to ensure terminology consistency across the workflow.

Related Tags