PROMPT ENGINEER
Vrinda International
- Location
- Remote (Remote, Maharashtra)
- Compensation
- $180k - $180k/yr
- Employment
- Full-time
- Level
- Senior Level
Posted 1 day ago
About the Role
Vrinda International is seeking a Prompt Engineer to design, test, and optimize prompts for LLM-based applications. The role focuses on improving AI output quality, reliability, and efficiency through structured experimentation and collaboration with engineering teams.
Skills
Prompt Engineering
LLM
Prompt Design
Prompt Optimization
Prompt Evaluation
Zero-shot Prompting
Few-shot Prompting
Prompt Chaining
RAG
Tokenization
Context Management
Hallucination Mitigation
Prompt Injection Prevention
AI Development
Perks
- Remote Work
Full job details
? HIRING | PROMPT ENGINEER ?
? Remote | ? 1 Position
? Experience: 9+ Years
? Budget: ₹1.8 LPM
? Contract | ⏳ Initial Duration: 3 Months
? Working Hours: 10 AM–7 PM IST
? ROLE OVERVIEW
Looking for an experienced Prompt Engineer with strong hands-on expertise in designing, testing, evaluating and optimizing prompts for LLM-based applications. The role focuses on improving AI output quality, reliability, consistency and efficiency through structured experimentation.
? MANDATORY
• 9+ years overall experience with relevant AI/LLM projects
• Formal Prompt Engineering Certification – Mandatory
• Hands-on LLM prompt engineering experience
• Prompt design, optimization & evaluation
• Zero-shot & few-shot prompting
• Prompt chaining & RAG
• LLM evaluation methodologies
• Context management, tokenization & context windows
• Identify/mitigate hallucinations, inconsistency & prompt injection
• Strong analytical and experimental mindset
? RESPONSIBILITIES
• Design and optimize prompts for AI use cases
• Develop reusable prompt templates/frameworks
• Run structured experiments and compare prompt variants
• Define evaluation metrics and analyze LLM responses
• Document experiments, results and recommendations
• Work with AI/Engineering teams to integrate optimized prompts
• Deliver measurable outcomes supporting go/no-go decisions