88Grounded knowledge chunks across resume, projects, skills, writing and impact evidence.
2Visitor journeys supported: recruiter evaluation and client project discovery.
1Consistent answer style shared by the floating assistant and dedicated AI Lab demo.
LiveInteractive assistant reachable from the homepage, AI Lab and project pages.
The Business Problem
A portfolio can contain strong evidence and still fail if visitors cannot find the right proof quickly. Recruiters may want to know whether I can do the job. A client may want to know whether I can build a BI platform, AI assistant or analytics dashboard. A technical reviewer may want to inspect the demos and delivery logic.
Ask Pharaoh AI turns that discovery problem into a conversational interface. Instead of asking visitors to read every section, the assistant helps them ask the question they actually came with.
The Product Question
I designed the assistant around one practical question:
Can a visitor ask about my work and receive a concise, evidence-grounded answer that points them toward the right proof?
This is different from a generic chatbot. The assistant is not trying to sound impressive. It retrieves the right evidence, structures it clearly and avoids claims that are not supported by my work.
My Role
I designed the assistant concept, wrote the knowledge-base chunks, structured the retrieval logic, refined response formatting, built the floating and dedicated chat interfaces, added sample questions, improved mobile behavior and shaped the answer style around recruiter and client decision-making.
The project sits at the intersection of AI development, product thinking, content design, search intent and portfolio strategy.
The Solution Architecture
The assistant works as a small evidence system. The main value is not the chat box itself; it is the structure behind the answers.
Portfolio evidence
Prepare the knowledge base Resume facts, project descriptions, impact metrics, skills, writing and career context are turned into retrievable chunks.
Grounded source layer Answers start from known evidence.
Question intent
Classify what the visitor needs Hiring fit, BI capability, AI projects, career history, certifications and stakeholder questions are routed toward relevant evidence.
Relevant retrieval The assistant selects context based on the question.
Answer design
Structure the response Answers are broken into direct response, evidence and next step, with bullet points where that helps scanning.
Readable output The visitor gets a clear answer without a wall of text.
Conversion path
Point to proof The assistant can direct visitors to demos, case studies, impact evidence, LinkedIn, GitHub or contact.
Next action The answer helps the visitor move forward.
Why It Matters
This assistant solves a portfolio problem that many technical professionals have: the work is broad, but the visitor has a narrow question. A recruiter may not care about every project. A client may not care about the career timeline. A technical reviewer may not care about the contact form until they see working proof.
Ask Pharaoh AI helps each visitor take a shorter path through the site while still keeping the answer grounded in the same evidence.
Response Design
The answer style is intentionally concise. It avoids long biography paragraphs and uses short sections such as direct answer, best fit, evidence and next step.
For hiring questions, the assistant is designed to be useful without sounding inflated. It explains where I am strongest, the type of work I can support and the evidence that makes the claim credible.
Grounding And Boundaries
The assistant is grounded in curated profile and portfolio chunks. If a question is outside the available evidence, the expected behavior is to answer cautiously and guide the visitor toward available proof instead of inventing unsupported details.
This is important because the assistant represents my professional brand. A confident unsupported answer would be less useful than a smaller grounded answer.
Use Cases
Recruiters
Recruiters can ask whether I am suitable for a role, what sectors I have worked in, what tools I use, what projects show evidence and how my background maps to a BI, analytics or AI role.
Clients
Clients can ask what type of analytics, dashboard, automation or AI product I can help build, then move directly to a demo or contact path.
Technical Reviewers
Technical reviewers can ask which projects demonstrate RAG, BI platforms, predictive analytics, document intelligence or dashboard design.
What The Project Does Not Prove
The assistant is not a general-purpose search engine and should not be treated as a verified external background check. I built it around curated evidence, retrieval, structured answer formatting and a visitor-friendly interface.
The value is in how the assistant reduces friction and makes proof easier to access.
Technology
Knowledge design: curated profile, project, writing and impact chunks. AI pattern: retrieval-grounded answer construction with explicit formatting rules. Interface: floating assistant plus full-page AI Lab demo. Delivery: static Cloudflare-ready frontend with a serverless AI route for enhanced responses where configured.
The Value In One Sentence
Ask Pharaoh AI Assistant shows how I design practical AI products that reduce discovery friction, retrieve evidence and help users act on information instead of making them read everything manually.