MCP Server Engineer

EXL; Gurugram District, Bengaluru; On-site; INR 1,800,000 - 3,000,000; Full time

Job summary

EXL is seeking a skilled MCP Server Engineer to design, build, and maintain secure connectors between Databricks workspaces and MCP servers in Gurgaon or Bengaluru. The role emphasizes low-latency data access for AI/ML workflows and strict governance.

Required hands-on experience with Python or TypeScript/Node.js SDKs, Databricks Delta Lake, Unity Catalog, Databricks SQL, and MLflow, plus knowledge of SSE, WebSockets, and JSON-RPC 2.0.

Qualifications

• Hands-on experience building MCP servers with Python or TypeScript/Node.js SDKs.
• Production experience with Databricks Delta Lake, Unity Catalog, Databricks SQL, and MLflow.

Responsibilities

• Design, build, and maintain secure connectors and pipelines between Databricks workspaces and MCP servers.
• Implement and configure MCP servers/clients to expose Databricks data securely to AI models and LLM applications.
• Optimize data retrieval, caching, and query performance between Databricks and LLM orchestration frameworks.
• Ensure data exposed through the MCP server adheres to enterprise data governance and Unity Catalog permissions.
• Collaborate with AI/ML engineers, data scientists, and software architects to define context, tools, and prompts for querying Databricks.
• Establish robust logging, error handling, and monitoring for high availability.

Skills

Python

TypeScript/Node.js

Databricks

LLM integration

API design

Tools

Databricks Delta Lake

Unity Catalog

Databricks SQL

MLflow

Job description

Role: MCP Server Engineer

Experience: 5+ Years

Location: Gurgaon OR Bangalore

Work Mode: Work From Office [5 Days Office]

ROLES AND RESPONSIBILITIES: Design, build, and maintain secure connectors and pipelines between Databricks workspaces (Delta Lake, Unity Catalog) and Model Context Protocol (MCP) servers. Implement and configure MCP servers/clients to expose Databricks data, schemas, and analytical tools securely to AI models and LLM applications. Optimize data retrieval, caching mechanisms, and query performance between Databricks and LLM orchestration frameworks to minimize latency. Ensure all data exposed through the MCP server adheres to strict enterprise data governance, access controls, and Unity Catalog permissions. Partner with AI/ML engineers, data scientists, and software architects to define the context, tools, and prompts required for LLM applications to effectively query Databricks. Establish robust logging, error-handling, and monitoring for the Databricks-MCP middleware to ensure high availability and reliability.

REQUIRED QUALIFICATIONS Proven, hands-on experience building, configuring, or extending MCP servers (using Python or TypeScript/Node.js SDKs) to connect LLMs to external data sources. Deep production experience with Databricks (Delta Lake, Unity Catalog, Databricks SQL, and Managed MLflow Strong understanding of SSE (Server-Sent Events), WebSockets, and JSON-RPC 2.0 protocols, which underpin MCP communication