Experience:
- 6–10+ years in software engineering
- 4+ years in GenAI / NLP / LLM-based systems
- Proven experience building production-grade AI systems
Key Responsibilities
- Design and build RAG pipelines (ingestion → indexing → retrieval → generation)
- Develop data ingestion pipelines:
- Document upload (PDF, DOCX, PPT)
- Parsing, table extraction, image extraction
- Image captioning & semantic chunking
- Metadata tagging & versioning
- Implement hybrid search (vector + keyword + optional graph)
- Build APIs using FastAPI (Python) for chat and orchestration
- Integrate LLMs (OpenAI / Claude via AWS Bedrock)
- Optimize response accuracy, latency, and cost
- Implement evaluation & observability (RAGAS, Langfuse/LangSmith)
Tech Stack
- Languages/Backend: Python, FastAPI
- LLM & Frameworks: LangChain, LangGraph, CrewAI
- Cloud: AWS (S3, ECS/Fargate, Lambda, API Gateway, Bedrock)
- Databases: Any Vector DB
- Data Ingestion: Unstructured, PyMuPDF, Tesseract, custom pipelines
- DevOps: Docker, GitHub Actions
- Observability/Eval: Langfuse, LangSmith, RAGAS, DeepEval
Requirements
- 6+ years in software engineering, strong in Python
- Hands-on experience with LLMs & RAG systems
- Experience building data ingestion pipelines and search systems
- Good understanding of embeddings, semantic search, and prompt engineering
- Experience with cloud deployment (AWS preferred)
Nice to Have
- Multi-modal RAG (images, tables)
- Agentic AI workflows (LangGraph / ReAct patterns)
- Experience with enterprise AI systems
