We're seeking a Prompt Engineer to join our team. This is your chance to apply your skills to a project with real-world impact.
Responsibilities:
- System Prompt Architecture: Design complex, structured system prompts. Use Chain-of-Thought and Few-Shot techniques to solve complex logic tasks (e.g., "Diagnose this car issue based on the user's description")
- Structured Generation (JSON): Engineer prompts that reliably output valid JSON for API integrations. You must ensure the model follows schema constraints 100% of the time
- Voice Persona Design: Control the agent's tone and pacing. Use instructions to influence output, making the agent sound empathetic, professional, or urgent as needed
- Jinja2 Integration: Embed dynamic data into prompts. You must write prompts that are robust enough to handle empty variables or unexpected data formats ({{ user_history }})
- Optimization: Continuously A/B test prompts to reduce hallucinations and improve conversion rates (e.g., "Did the agent successfully book the meeting?")
Requirements:
- Deep LLM Expertise: You know the difference between GPT-4 and Claude 3 Opus. You understand token limits, temperature, and context windows
- Technical Proficiency: You are comfortable working with JSON and Markdown. You can write basic Jinja2 templates
- Analytical Mindset: You don't just "guess" improvements; you analyze conversation logs to find the root cause of errors
- Linguistic Precision: Excellent command of English to catch subtle nuances in interaction
- Language: Russian (Native/Fluent) and English (B2+)
Contact us:
Показать контакты
Будьте осторожны: если работодатель просит войти в их систему, используя iCloud/Google, прислать код/пароль, запустить код/ПО, не делайте этого - это мошенники. Обязательно жмите "Пожаловаться" или пишите в поддержку. Подробнее в гайде →
Текст вакансии взят без изменений
Источник - Telegram канал. Название доступно после авторизации