RAG AI Developer (LLM + Retrieval) – EdTech

AP Guru; Mumbai; On-site; INR 1,100,000 - 1,700,000; Full time

Job summary

AP Guru is seeking an experienced RAG AI Developer to build and improve AI features for EdTech products, including course Q&A bots, tutor assistants, content search and internal knowledge assistants. You will handle document ingestion, embeddings, retrieval pipelines, evaluation and deployment.

You will work with providers of vector databases, implement hybrid search, and ensure efficient, scalable APIs with monitoring and testing.

Qualifications

• 1+ year experience building NLP/LLM features (hands-on RAG or retrieval work).
• Experience with embeddings, chunking strategies, and document loaders (PDF/HTML/Doc).
• Familiarity with at least one vector DB and retrieval methods (cosine similarity, MMR).
• Understanding of basic ML concepts and text preprocessing.

Responsibilities

• Implement hybrid search (semantic + keyword), reranking, filters, and metadata-based retrieval.
• Integrate LLMs with tools/frameworks (e.g., LangChain / LlamaIndex or custom pipelines).
• Work with vector databases (Pinecone, Weaviate, FAISS, Chroma, Milvus) and optimize retrieval performance.
• Create evaluation metrics for RAG quality (faithfulness, relevance, context precision/recall) and reduce hallucinations.
• Build prompt templates, guardrails, and citation-based answers.
• Deploy services/APIs (FastAPI/Flask), monitor latency/cost, and implement caching strategies.
• Collaborate with product/content teams to define data sources and user workflows.

Skills

NLP/LLM features

Embeddings

Document loaders

ML concepts

Tools

Pinecone

Weaviate

FAISS

Chroma

Milvus

Job description

We are looking for a RAG (Retrieval-Augmented Generation) AI Developer to build and improve AI features for our EdTech products—such as course Q&A bots, tutor assistants, content search, and internal knowledge assistants. You will work on document ingestion, embeddings, retrieval pipelines, evaluation, and deployment.

Key Responsibilities:

• Implement hybrid search (semantic + keyword), reranking, filters, and metadata-based retrieval.
• Integrate LLMs with tools/frameworks (e.g., LangChain / LlamaIndex or custom pipelines).
• Work with vector databases (e.g., Pinecone, Weaviate, FAISS, Chroma, Milvus) and optimize retrieval performance.
• Create evaluation metrics for RAG quality (faithfulness, relevance, context precision/recall) and reduce hallucinations.
• Build prompt templates, guardrails, and citation-based answers.
• Deploy services/APIs (FastAPI/Flask), monitor latency/cost, and implement caching strategies.
• Collaborate with product/content teams to define data sources and user workflows.

Required Skills & Qualifications:

• 1+ year experience building NLP/LLM features (must have some hands-on RAG or retrieval work).
• Experience with embeddings, chunking strategies, and document loaders (PDF/HTML/Doc).
• Familiarity with at least one vector DB and retrieval methods (cosine similarity, MMR, etc.).
• Understanding of basic ML concepts and text preprocessing.

Preferred (Nice to Have):

• Experience with OpenAI / Anthropic / Google / open-source LLMs (Llama, Mistral, etc.).
• Experience with OCR pipelines (for scanned PDFs), speech/text, or multilingual content (helpful for EdTech).
• Experience with Docker, cloud deployment (AWS/GCP/Azure), CI/CD.
• Prior work on chatbots, tutoring systems, or knowledge bases.

What Success Looks Like (KPIs):

• Higher answer accuracy + lower hallucination rate
• Faster retrieval latency and lower compute cost
• Clear citations and better user satisfaction on Q&A flows