Hey, I'm
Keenan
Carroll
I design and build AI-powered systems that automate real-world business operations — bridging the gap between AI capability and real-world execution.
Chat with my AI ↓
█
Chat with my AI ↓
█
About Me
I design and build AI-powered systems that solve real business problems — from support automation and document intelligence to multi-agent platforms that replace entire workflows. My focus is on production-grade systems that work, not experiments that demo well.
I'm currently transitioning into AI engineering after years in enterprise SaaS — first as a Tier III Technical Engineer, then as an Enablement Manager at WalkMe, now acquired by SAP. That background gives me something most AI engineers don't have: I understand how enterprise software actually gets used, where it breaks, and what teams need from it.
My goal is to build AI systems that close the gap between what's technically possible and what businesses can actually use.
AI Agents
Multi-agent systems coordinating tasks across platforms
Automation
Intelligent communication and workflow pipelines
Data Systems
Structuring unorganized data into usable infrastructure
Operations
Reducing manual work in operations-heavy businesses
What I've Built
Apex CRM Sales Intelligence
RAG-Powered Competitive Intelligence Agent
A hybrid RAG sales assistant built with LangChain and Claude. Answers competitive questions by searching internal battlecards and playbooks, live web data, or both — depending on what the question needs. Uses a ReAct agent to reason step-by-step and pick the right tool for each query. Includes an LLM-as-judge eval suite scoring answers on relevance, groundedness, and completeness.
A purpose-built command center for a real working rock band — replacing scattered Google Docs, spreadsheets, and group chats with a single integrated system. Features 12 AI personas (band manager, publicist, booking agent, finance specialist, and more) with distinct personalities and live write access to the database. Syncs from Google Calendar, Spotify, Square, YouTube, Instagram, and Discord. Includes a Felix Discord bot that consults the full AI team mid-conversation.
An AI research assistant that routes natural language queries to internal documents or live web search — automatically. Built on the ReAct agent pattern with a visual reasoning trace so you can follow every thought, action, and observation in real time. Combines Voyage AI semantic search over a ChromaDB knowledge base with Tavily web retrieval in a single query.
A support tool that answers questions by searching a knowledge base of PDF documents. Loads docs into a vector database, finds relevant chunks via semantic search, and generates concise answers using Claude — with full source attribution. Built with LangChain, Chroma, Voyage AI embeddings, and Streamlit.
A production-style Retrieval Augmented Generation system that lets users upload a PDF and ask natural language questions — getting answers grounded exclusively in their document. No hallucination, no out-of-scope responses. Demonstrates the RAG pattern powering the next generation of B2B AI features.
Where I've Been
Independent AI Builder & Product Developer
Self
Design and ship production-grade AI systems — RAG pipelines, LangChain agents, and multi-agent platforms deployed to real users. Focus on systems that work in the real world, not just demos.
- ▸Built Apex CRM Sales Intelligence — a hybrid RAG agent with LLM-as-judge evals, deployed on Railway
- ▸Shipped Valletta Command Center — a full-stack platform with 12 AI personas and live third-party integrations
- ▸Developed multiple RAG pipelines using LangChain, Claude, Voyage AI, and ChromaDB
Enablement Manager
Technology Sector
Manages internal and external enablement programs in a technology environment. Developing systems thinking and cross-functional coordination skills that directly inform how I design AI workflows today.
- ▸Cross-functional program management and stakeholder alignment
- ▸Systems design and process documentation at scale
- ▸Technology adoption and training program delivery
Tier III Technical Engineer
Technology Sector
Handled complex escalations and deep technical troubleshooting as the final tier of support. Built a strong foundation in diagnosing systems at the code and infrastructure level — experience that directly shapes how I approach building and debugging AI systems today.
- ▸Owned resolution of high-complexity technical escalations
- ▸Developed deep product knowledge across multiple system layers
- ▸Collaborated cross-functionally with engineering to surface and resolve recurring issues