Posted on: 28/11/2025
Description : Prompt Engineer.
Experience : 3-5 years.
Location : Hyderabad (Onsite/Hybrid).
Job Description :
We are looking for a Prompt Engineer with strong software engineering expertise and hands-on experience working with Large Language Models (LLMs) and Agentic AI frameworks.
The ideal candidate will design, test, and optimize prompts, retrieval pipelines, and agentic workflows that power intelligent, production-grade AI systems.
Key Responsibilities :
- Design, prototype, and refine prompts for LLMs (OpenAI, Anthropic, Gemini, Llama, etc.)
- Build and integrate LLM-based services via APIs with FastAPI (routers, DI, auth, rate limiting).
- Implement Retrieval-Augmented Generation (RAG) workflows document parsing, chunking, embeddings, vector search, and hybrid retrieval.
- Develop agentic flows and state machines using LangGraph; build chains/tools using LangChain.
- Create and maintain testing frameworks for prompts, retrieval logic, and chain/tool performance (A/B testing, relevance, latency, and cost).
- Manage vector database integrations (FAISS, Pinecone) and ensure retrieval accuracy.
- Collaborate on end-to-end API development, CI/CD integration, and system monitoring.
- Maintain documentation, best practices, and advocate for prompt engineering standards across teams.
Must-Have Skills :
- 3-5 years of software engineering experience with strong Python skills.
- Proven hands-on experience with FastAPI, LLMs (OpenAI, Anthropic, Gemini, Llama), and function/tool calling.
- Practical experience implementing RAG, embeddings, hybrid retrieval, and document chunking.
- Proficiency with LangGraph and/or LangChain for agentic and chained workflows.
- Experience with vector databases (FAISS, Pinecone, Weaviate) and document preprocessing.
- Familiarity with Git, Docker, and CI/CD pipelines.
- Strong communication and collaboration skills.
Nice-to-Have Skills :
- Experience in LLM fine-tuning (LoRA, quantization) and monitoring prompt performance.
- Knowledge of cloud deployment (AWS/Azure/GCP) and LLM observability (latency, cost tracking).
- Prior exposure to regulated domains such as healthcare or life sciences.
Did you find something suspicious?