Skip to main content
redennInformatics
All open roles

AI Engineer (LLM Applications)

  • Patiala, India · On-site
  • 2 – 5 years
  • Full-time
  • Python
  • Claude & OpenAI APIs
  • RAG
  • TypeScript
  • pgvector / Postgres
  • Prompt engineering
Apply for this role

You will own the LLM systems Redenn already runs in production: the sales chatbot that talks to visitors on client sites across six countries, the retrieval layer that keeps it grounded in each client's real services and pricing, and the agent workflows inside the Pro Lifeset CRM. New work lands every month, from document extraction for a clinic to an AI assistant sitting inside an ecommerce build.

Why this role

Most AI jobs in this market are pilots and demos that never ship, while here the prompt you tune on Tuesday is answering a real customer on Wednesday and there is nobody between you and the founder to slow that down.

A normal week

  1. You rewrite a system prompt, run it against last week's real conversations, and push it to a live client site the same afternoon.
  2. You sit with the SEO and sales team to find the questions visitors keep asking that the bot still answers badly.
  3. You wire a new retrieval index over a client's service pages, PDFs and pricing sheets, then deliberately try to break it before a customer does.
  4. You pair with a backend developer so an agent workflow can read and write CRM records safely.

What you’ll do

  • Design and maintain RAG pipelines end to end: chunking, embeddings, retrieval, grounding and citations.
  • Ship chatbot and agent features on the Claude and OpenAI APIs, including tool use, streaming and structured output.
  • Build evals and regression checks so a prompt change never quietly breaks a live bot.
  • Watch cost, latency and token usage in production and bring them down.
  • Integrate AI features into the existing Next.js and Postgres stack alongside the web team.

What we’re looking for

  • 2 to 5 years writing production software, including at least one LLM feature you took all the way to real users.
  • Strong Python or TypeScript, with real comfort around async work, queues and API design.
  • Hands-on with an LLM API: prompting, tool calling, streaming, structured outputs and their failure modes.
  • Working knowledge of embeddings and vector search, and why retrieval quality decides chatbot quality.
  • Solid SQL and a real understanding of the data you are retrieving from.

Nice to have (not required)

  • You have used pgvector or Supabase for vector search, though Pinecone or Chroma experience transfers fine.
  • Some exposure to evals or LLM observability, even if it was homemade logging and a spreadsheet.
  • Enough React to build the small UI your model feature needs without waiting on anyone.

Sound like you?

Send your CV and a short note. A real person reads every application — expect a reply within a few business days.

Apply for this role