Prompt Engineer - Fully Remote - India

Halo believes in innovation by inclusion to solve digital problems. As an international agency of over 200 people specializing in interactive media strategy and development, we embrace equity and empowerment in a serious way. Our interdisciplinary teams of unique designers, developers and entrepreneurial minds with a variety of backgrounds, viewpoints, and skills connect to solve business challenges of every shape and size. We empathize to form deep, meaningful relationships with our clients so they can do the same with their audience. Working at Halo feels like belonging. Learn more about our philosophy, benefits, and team at https://halopowered.com/
Role overview : Prompt Engineer to join the team. The ideal candidate will have a RAG based experience optimizing prompts for 4.o.

Key Responsibilities:

• Design, develop, and refine prompts to improve the quality and relevance of AI outputs.
• Conduct experiments to evaluate prompt effectiveness and iterate based on results.
• Analyze user interactions and feedback to inform prompt adjustments and enhancements.
• Widely used for AI, prompt engineering, RAG pipelines, and LLM orchestration
• Useful if the team integrates prompts into web-based interfaces or uses Node.js APIs.
• Used for querying metadata or integrating with structured databases behind RAG systems

Requirements

• Proficiency working with 4.o in RAG applications.
• Excellent analytical and problem-solving abilities.
• Strong communication skills to convey complex concepts clearly.
• Familiarity with user experience design principles is a plus.
• Python
1. Building and managing prompt workflows and pipelines.
2. Familiarity with LangChain, LlamaIndex, or similar libraries.
3. Working with OpenAI SDK (e.g., openai.ChatCompletion.create()).
4. Data parsing and transformation for building context windows.
5. Writing evaluation scripts for prompt performance (A/B testing, token cost analysis).
6. Experience with async I/O (for parallel prompt testing or streaming responses).
• JavaScript/TypeScript (if used)
1. Frontend experimentation or tooling for UX-related prompt testing.
2. API integration for frontend-driven RAG queries.
3. Using frameworks like Next.js or React (if applicable to the product’s delivery mechanism).
• SQL
1. Writing queries to extract structured user data or feedback logs.
2. Possibly used in hybrid RAG systems combining structured (DB) + unstructured sources.
• Frameworks & Tools
1. RAG Frameworks - LangChain, LlamaIndex, Haystack
2. LLM APIs - OpenAI GPT-4o, Azure OpenAI API, Anthropic Claude API
3. Vector Databases - Pinecone, Weaviate, FAISS, Chroma, Milvus
4. Embeddings - OpenAI Embeddings, SentenceTransformers (HuggingFace)
5. Data Stores / DBs - PostgreSQL, MongoDB, Redis (for caching context/state)
6. Evaluation & Experimentation - Prompt injection testing, prompt tuning, cost evaluation, latency analysis
7. API Integration - Prompt injection testing, prompt tuning, cost evaluation, latency analysis
8. UX/Data-Driven Prompting - Understanding of user intent analysis, heatmaps, feedback loops for prompt refinement
9. CI/CD & Tooling - Git, GitHub Actions, Docker (for deploying prompt pipelines or microservices)
10. Software Engineering Concepts - Modularity, version control, testing (unit & integration), observability/logging practices

Benefits

• 100% RemoteWork.
• Salary in USD.
• Get to work on challenging projects for the U.S