About Lifelong :

Since 2015, Lifelong Online has been redefining everyday living with innovative, reliable, and thoughtfully designed products. With a portfolio of over 300 products, we've become a trusted part of10 million households across India, delivering to over 10,000 pin codes. Our offerings span two keyverticals: Sports, Fitness, and Personal Grooming, where we empower active lifestyles and self-care, and Home Solutions, encompassing kitchen and home essentials, home improvement tools, electronics, and baby care. At Lifelong Online, we're committed to making life simpler, smarter, and more enjoyablecrafted for you and proudly made for India.

What You'll Work On :

- Architect and ship production agentic AI systems for the operations platform - multi-step workflows, tool-use, decision logic with safety guardrails.

- Build and maintain RAG pipelines over enterprise data - operational logs, business reports, partner-channel data, and internal knowledge bases.

- Design and integrate with Model Context Protocol (MCP) servers to expose internal tools to AI agents.

- Own multi-model LLM orchestration - route between leading commercial and open-source LLMs on Bedrock and Vertex AI based on cost, latency, and capability.

- Build evaluation harnesses - RAGAS, prompt regression, golden-set eval pipelines - that gate AI feature releases.

- Implement responsible AI practices - prompt injection defense, OWASP LLM Top 10 coverage, data leakage validation, bias audits, audit trails.

- Partner with the QA engineer on AI eval automation in CI/CD.

- Surface AI decisions through APIs that the full-stack team integrates into operator dashboards.

Who Should Apply :

- 4+ years of professional engineering experience, with at least 1.5 years building production AI/LLM systems.

- Strong Python expertise - FastAPI, async patterns, Pydantic, production-grade code.

- Hands-on with at least one LLM ecosystem in production : LangChain/LangGraph, LlamaIndex, Haystack, or equivalent.

- RAG production experience - chunking strategies, embedding models, vector DBs (Qdrant, Pinecone, Weaviate, pgvector), hybrid retrieval.

- Experience with multi-model deployment - at least one of AWS Bedrock, Google Vertex AI, Azure OpenAI in production.

- Strong understanding of LLM evaluation - beyond accuracy. You know what RAGAS metrics mean and how to set up eval harnesses.

- Experience with Model Context Protocol (MCP) - building or integrating MCP servers.

- Exposure to agentic frameworks (CrewAI, AutoGen, LangGraph) in production.

- Bonus : Background in retail / D2C / consumer / supply chain - bonus if you've worked on operational AI.

- Bonus : Open-source contributions in the AI/ML space.

What We Offer :

- End-to-end ownership of the AI layer - architecture, implementation, and measurable operational impact.

- Modern AI infrastructure - Bedrock and Vertex AI in production, with MCP as a core platform layer.