Agentic AI jobs
Position: Agentic AI Developer
Location: Research Triangle Park (RTP), North Carolina – Onsite
Duration: Long Term Contract
Project: AI Analytics Platform
About the Role:
We are building an enterprise-grade AI-powered analytics platform that enables business users to query complex datasets through natural language. The platform uses a multi-agent architecture where specialized AI agents collaborate to route questions, generate SQL, execute queries against Snowflake, produce insights, and create dynamic visualizations – all autonomously.
We are looking for consultants who can contribute to the development of new domain agents, extend the platform’s capabilities, and help scale the architecture to support additional business domains.
What You Will Work On
Develop and integrate new domain-specific AI agents into an existing multi-agent orchestration system built on LangGraph and LangChain
Design and implement SQL generation agents that translate natural language questions into precise Snowflake SQL, enforcing business rules, RBAC, and fiscal period logic
Build and consume MCP (Model Context Protocol) server integrations for secure, structured data access across enterprise data sources
Work with PromptQL and RAG/PageIndex patterns to improve query accuracy, context retrieval, and domain-specific grounding
Develop FastAPI endpoints and async workflows to support real-time query processing, background job execution, and frontend integration
Create intelligent data visualization pipelines that automatically select and generate the right chart type (heatmaps, bar charts, KPIs) based on query results and user intent
Write domain context configurations (JSON schemas) that define column mappings, business rules, valid values, and metric definitions for each data domain
Contribute to a plugin-based domain registry architecture that allows new agents to be added without modifying core orchestration code.
Required Technical Skills
Python – Expert-level proficiency; ability to write production-quality async code, work with complex class hierarchies, and debug multi-layer systems
Snowflake – Strong experience writing and optimizing analytical SQL; understanding of RBAC patterns, fiscal period logic, and aggregation queries
LangGraph / LangChain – Hands-on experience building stateful, multi-step AI agent workflows with tool calling, checkpointing, and conditional routing
RAG / PageIndex – Practical experience implementing retrieval-augmented generation pipelines for context-aware AI applications
MCP (Model Context Protocol) – Practical experience with MCP client/server patterns for structured tool-to-data communication
Snowflake Cortex – Experience with Snowflake Cortex AI functions (COMPLETE, EXTRACT_ANSWER, SUMMARIZE, SENTIMENT) and building AI/ML workflows natively within Snowflake
PromptQL – Experience with structured prompt engineering and query language patterns for grounding LLM outputs in enterprise data
Preferred Skills
Experience with Plotly for programmatic chart generation
Pandas for data transformation and pivot operations
Docker and containerized deployment workflows
Enterprise authentication patterns (OAuth 2.0, token management)
Experience working within large enterprise codebases with multiple contributors
Minimum Qualifications
5+ years of professional Python development experience
2+ years building AI/ML or LLM-powered applications in production
Demonstrated experience with at least two of: LangChain, LangGraph, PromptQL, or equivalent agent orchestration frameworks, MCP
Strong SQL skills with experience on Snowflake and snowflake cortex
Ability to work independently, understand existing architecture quickly, and deliver production-ready code with minimal supervision
To apply for this job email your details to srujan@sourcemantra.com