AI Integration Engineer
0 views2-5 years
Posted today
- AWS
- Docker
- Kubernetes
- Llm
- PostgreSQL
- Python
- React
- TypeScript
AI Integration Engineer (AI Engineer)
Company: AJEMS
Job location: Pune, Maharashtra (work from office)
Type: Full-time ·
Experience: 2 to 5 years
About AJEMS
• AJEMS is an enterprise platform where businesses build their own apps, forms and workflows without code, then get answers from that data. Brain Behind Business.
• AJEN is the AI assistant inside the product. Users ask questions in plain language and get reports, insights and answers from their own workspace.
• Connectors bring data in from Tally, Google Sheets, Excel, Meta Lead Ads and more, so everything sits in one place.
The role
You own how AI works inside AJEMS: AJEN, the AI layer behind it, and how outside AI tools like Claude and ChatGPT connect to a customer's workspace. It runs on real customer data, so getting security right matters as much as getting answers right.
What you will do
• Build and improve AJEN: chat over a customer's apps, forms and records, plus AI insights, reports and template generation
• Design tool calling so the model queries live workspace data instead of guessing
• Extend the AJEMS MCP server so Claude, ChatGPT and other AI clients can read and write workspace data securely
• Build an AI gateway: model routing, PII filtering, logging, cost and rate limits, so switching model providers is a config change
• Test and run open-source models (Qwen, Llama, DeepSeek) through hosted APIs and on our own GPU servers
• Enforce strict tenant isolation: the AI only ever sees data from the tenant asking
• Move all AI features to a secure backend service
• Build evals and test sets that measure answer accuracy, hallucination and latency
• Use AI inside connectors for field mapping, data cleaning and schema detection when importing Excel, Sheets or Tally data
• Work closely with the React and backend developers and the QA team
Must have
• 2+ years building backend software, with at least 1 year shipping LLM features to real users
• Strong Python, working knowledge of Node.js or TypeScript
• Hands-on with LLM APIs (OpenAI, Anthropic or open-source), tool calling and structured JSON output
• RAG basics: embeddings, a vector store (pgvector, Qdrant or similar), chunking and retrieval tuning
• REST APIs, OAuth 2.0, API keys, JWT and webhooks
• Comfortable with PostgreSQL and SQL, including safe text-to-SQL or query building
• A security mindset: prompt injection, data leakage and PII handling
• Uses AI coding tools daily and can explain every line shipped
Good to have
• Built an MCP (Model Context Protocol) server or client
• Self-hosted models with Ollama or vLLM, quantisation, GPU setup
• Multi-tenant data models
• Django and React
• Agent frameworks (LangGraph or similar), and knowing when not to use one
• Docker, Kubernetes, AWS
• APIs for Tally, Zoho, Google Workspace or Meta
Why join
• Build the AI layer of a product that launched in November 2026
• Make core architecture decisions early, not maintain someone else's
• Small team, direct ownership, real users from day one
How to apply
Share your resumes at admin@ajems.com