About the role
About the role
LangChain · PostgreSQL · Python · Multi-LLM
We are building the next generation of AI-powered products at D2R Tech. We need engineers who can go beyond prompt engineering people who architect multi-agent systems, integrate LLMs intelligently into production backends and own the full AI pipeline from retrieval to response. If you have shipped LangChain applications, designed RAG pipelines or worked across multiple LLM providers (OpenAI, Anthropic, Gemini, Mistral, open-source), this role is for you.
What you will do
Design and build scalable data pipelines for audit data
Implement real-time analytics for compliance dashboards
Optimize data storage and query performance
Build data quality and lineage monitoring systems
Requirements
LangChain & AI Frameworks
• Strong hands-on LangChain experience in LCEL, chains, agents, memory, tools
• LangGraph for stateful multi-agent workflows
• Familiarity with LangSmith or similar observability/tracing tools
• Experience with prompt templating, output parsers, and structured outputs
Multi-LLM Proficiency
• Production experience with 2+ LLM providers: OpenAI GPT-4/4o, Anthropic Claude, Gemini, Mistral
• Understanding of model selection trade-offs like cost, context window, capability
• Implementing LLM fallback, routing, and load balancing strategies
• Open-source model experience (LLaMA, Mistral, Phi) via Ollama or HuggingFace is a plus
PostgreSQL & Data
• Strong PostgreSQL skills like schema design, query optimization, indexing
• pgvector for vector similarity search and embedding storage
• Designing session and conversation memory schemas
• Experience with async DB access (SQLAlchemy async, asyncpg)
Python & Backend
• Proficient Python like async/await, Pydantic, clean architecture
• FastAPI for building AI-backed REST/streaming endpoints
• Docker for containerized development and deployment
• Redis for caching, rate limiting, and ephemeral session data