AI Engineer (Applied LLMs and Agents)

Plantnxt Technologies; Chennai; On-site; INR 1,800,000 - 3,000,000; Full time

Job summary

PlantNXT Technologies Pvt. Ltd. is seeking an AI Engineer to design and implement LLM-powered features and agent workflows on a multi-tenant Node.js/PostgreSQL platform. You will lead RAG pipelines, prompt design, and evaluation to ensure reliable, cost-efficient responses and safe human-in-the-loop escalation.

The role requires hands-on experience with LLM APIs, TS/Node.js, Python, and practical RAG expertise to deliver production-grade AI capabilities for manufacturing-focused decision support.

Qualifications

• 2+ years in software engineering, with LLM experience in production.
• Strong TypeScript/Node.js and Python skills.
• Hands-on with LLM APIs and prompts.
• Practical RAG experience and evaluation.
• Ability to assess hallucination risk and guardrails.
• Disciplined testing and regression checks.

Responsibilities

• Design and build LLM-powered features and agent workflows.
• Build retrieval systems (RAG) over documents and data.
• Define when to answer, cite, refuse or escalate to humans.
• Create golden datasets and automated evaluation for prompts and retrieval.
• Integrate AI features into Node.js and PostgreSQL with tenant isolation.
• Monitor cost, latency and answer quality in production.
• Collaborate with ML engineer on model outputs for the decision layer.

Skills

LLM in production

TypeScript/Node.js

Python

LLM APIs

RAG systems

Evaluation discipline

Guardrails

Cost monitoring

Human-in-the-loop

Tools

PostgreSQL pgvector

Job description

Job Description : AI Engineer (Applied LLMs and Agents)

Company: PlantNXT Technologies Pvt. Ltd.
Location: Chennai

About the role

PlantNxt is an AI Native Decision support Platform purpose-built for the manufacturing industry. It unifies plant data and delivers real-time, prescriptive decisions in natural language. The innovation lies in its vertical-specific design, pre-trained for manufacturing operations and ready from day one, helping leaders make decisions quicker that drives performance and profitability

What you will do

• Design and build LLM-powered features and agent workflows, using the Claude API and, where useful, local models
• Build retrieval systems (RAG): chunking, embeddings, indexing and retrieval evaluation over manufacturing documents and structured data
• Define when the system should answer, cite, refuse or elevate to a human
• Create golden datasets and automated evaluation for prompts, retrieval and agent behaviour
• Integrate AI features into a Node.js and PostgreSQL platform, respecting strict per-tenant data isolation
• Monitor cost, latency and answer quality in production, and improve them
• Work with the ML engineer on model outputs that feed the decision layer

Must have

• 2+ years of professional software engineering experience, with at least 1 year building LLM features in production
• Strong TypeScript or Node.js, plus working Python
• Hands-on experience with LLM APIs: prompt design, tool use, structured outputs
• Practical RAG experience, including retrieval evaluation, not only a demo
• A disciplined approach to evaluation: test sets, regression checks on prompt changes, and error analysis
• Clear reasoning about hallucination risk and guardrails in a domain where wrong answers matter

Nice to have

• Experience running local models (Ollama or similar) and choosing between local and hosted models on cost, latency or privacy grounds
• Background in operations, supply chain, engineering or manufacturing software
• Experience designing agents with human-in-the-loop approval steps
• Awareness of data isolation and security practices for enterprise customers
• Experience with PostgreSQL vector search (pgvector) or similar