Now Live
The Briefing Assistant
The first chatbot I shipped lives inside the Baylor × SAP pitch — a briefing assistant that knows every activation, every data point, and every talking point in the proposal. Executives can ask it questions, compare options, and prep for conversations, all within the site.
Context-Aware
Knows the full SAP pitch — jersey patch ROI, suite packages, keynote logistics, app features, measurement methodology.
Instant Answers
Answers questions in natural language without digging through slides. Ask about valuation, compare activation tiers, get ROI numbers.
Embedded in the Site
Accessible from any page in the SAP pitch via a persistent floating panel — no separate login or tool required.
Briefing Assistant
Ready
What's Coming
The full AI layer
Every part of this portfolio eventually gets an AI interface — projects, dashboards, social brands. Here's the build order.
SAP Briefing Assistant
Embedded in the Baylor × SAP pitch. Answers questions about all 7 pages of the proposal in natural language.
Portfolio-Wide Bot
A single AI that knows every project in this portfolio — ask it about dashboards, social strategy, class work, or my background.
Social Media AI
AI assistant trained on each brand's voice and strategy — generates captions, answers questions about content approach, and analyzes performance.
Dashboard AI Layer
Natural language interface for the personal and business dashboards — ask questions about your data instead of reading charts.
Recruiting Assistant
An AI that answers recruiter and employer questions about my experience, projects, and skills — available 24/7 directly from this portfolio.
More to come
Class Project • MIS 4V90
Where it started: the Python chatbot
Before the SAP Briefing Assistant, before the roadmap — there was a terminal window and a blinking cursor. The class project chatbot is where I first learned how language model APIs actually work: stateless calls, manual context management, and the power of a well-written system prompt.
Stack
Python 3.11 + Gemini API — no frameworks, no frontend. Just a while True loop and an API key.
Key Insight
LLM APIs are stateless — memory is always the developer's job. The fix: send the full conversation history on every single call.
System Prompt
Custom persona baked in at init — tone, constraints, and capabilities defined before the first message is ever sent.