AI/LLM Engineer

AI/LLM Engineer: Sales Automation & RAG Development

Duration: 12 Weeks (3 Sprints)

Timeline: Second week of February – Mid-April

Commitment: Part-time (2–3 hours daily availability required)

Location: Remote

The Mission

We are building an intelligent sales enablement tool designed to revolutionize how our team prepares for pitches. The goal is to develop an Agentic AI Assistant that guides a salesperson through a discovery process, autonomously researches prospects (via web and internal data), and automatically populates high-fidelity PowerPoint decks.

The Scope of Work

You will be the primary engineer responsible for architecting and deploying the following:

Agentic Chatbot Interface: Build a conversational flow that proactively asks specific questions to extract necessary context from the salesperson.

Hybrid RAG Pipeline: Develop a system that retrieves and synthesizes data from both the live web and internal sales repositories.

Automated Document Generation: Integrate the LLM output into a pre-defined PowerPoint template (PPTX) with high structural accuracy.

API & Deployment: Build the backend using FastAPI and ensure the system is production-ready.

Key Technical Requirements

To be successful in this role, you must have hands-on experience with:

Python Mastery: Expert-level Python for production-grade code.

Agent Frameworks: Proven experience with LangChain, LangGraph, or CrewAI (preference for LangGraph for complex state management).

RAG Architecture: Experience with vector databases, embedding models, and optimizing retrieval for accuracy.

Server Development: Building robust APIs using FastAPI.

PPTX Automation: Familiarity with libraries (like python-pptx) to programmatically manipulate slide templates.

DevOps: Experience taking an AI project from a local notebook to a production-deployed environment.

Project Timeline & Milestones

The project is structured into three distinct 4-week sprints:

Sprint 1 (February): Discovery logic, chatbot architecture, and initial internal data connection.

Sprint 2 (March): Web search integration and RAG pipeline optimization.

Sprint 3 (April): PowerPoint automation, testing, and production deployment.