Large Language Models jobs abroad
617 job offers found
explore Thessaloniki , Thessaloniki
Lead Prompt Engineer: Strategic LLM Migration open_in_new
explore Remote , United States
DESCRIPTION
- Strategic Oversight: Manage the full end-to-end technical migration workflow, ensuring a seamless transition from human-rated templates to automated LLM systems.
- Team Mentorship: Guide a team of engineers through the complexities of parent-child template clusters and automated optimization.
- Advanced Optimization: Direct the use of APG and APO tools, providing the high-level intervention needed when automated systems reach plateaus.
- Launch Certification: Take final responsibility for accuracy metrics, running prompt versions against gold data to verify $F_1$ scores, precision, and recall.
- Edge-Case Resolution: Architect manual prompt strategies to solve the most difficult "anti-patterns" in legacy architectures.
- Schedule: Part-Time (Flexible hours within project milestones)
- Location: 100% Remote (Must be based in the United States)
- Employment Type: Freelance / Independent Contractor
- Duration: Long-term technical migration project
- Elite Expertise: At least 7 years of experience as a Prompt Engineer with a proven history of tuning LLMs for structured outputs and complex classification.
- Advanced Academic Background: Master’s or Doctorate (PhD) in Computer Science, Computational Linguistics, HCI, or a related analytical field.
- Data Mastery: Strong proficiency in SQL and deep-dive error pattern analysis to monitor and enhance model performance.
- Technical Agility: Expert-level ability to master proprietary tools and enterprise-grade interfaces like the Goose API.
- Leadership Excellence: A natural mentor with the verbal and written communication skills required to justify technical launches to stakeholders.
- Experience in software engineering or AI model evaluation at an enterprise scale.
- Deep linguistic expertise, specifically in semantics and formal logic.
- Hands-on experience with Chain-of-Thought (CoT) and shadowbot disagreement tracking.
