Overview

An AI-powered conversational interface designed to explore a candidate’s professional background through natural language, rather than static documents.

Instead of presenting information linearly, the agent supports intent-driven exploration of experience, projects, and role fit, while remaining constrained, interpretable, and reliable.

Why this exists

Traditional resumes are static and linear by necessity.

This project explores whether a conversational interface can surface relevant experience more effectively by responding to how people actually ask questions, such as focusing on a specific role, skill set, or industry context, without overwhelming them with unrelated information.

The goal is not to replace resumes, but to examine how conversational AI might function as a complementary interface alongside them.

Conversational architecture

The agent is structured around a limited set of intentional conversation paths, including:

  • Professional experience overview

  • Role and skill fit, including questions based on job descriptions

  • Availability and work arrangements

Rather than supporting unrestricted dialogue, the interaction is designed to guide users through focused prompts and clarifying follow-ups. This keeps responses aligned with the underlying knowledge source and reduces ambiguity and overreach.

Guardrails and constraints

To maintain reliability, the agent is limited to a curated knowledge base derived from my CV and selected project summaries.

It does not speculate beyond this material or attempt to infer information that cannot be traced back to the source. When a question falls outside scope or lacks sufficient context, the agent makes that limitation explicit and redirects the conversation accordingly.

These constraints are intentional and reflect how I approach scope control in production conversational systems.

Tone of voice

The tone is deliberately professional, clear, and neutral.

The agent avoids playful or overly personalized language in favor of clarity and interpretability, prioritizing usefulness over personality. This reflects both the intended audience and the types of decision-making contexts in which conversational interfaces are often used.

Interaction design principles

  • Clarity over cleverness

  • Guidance over open-endedness

  • Reliability over expressiveness

The focus was not on making the agent feel impressive, but on designing an interaction that is useful, constrained, and human-centered rather than performative.

What this demonstrates

This project demonstrates my approach to conversational architecture, tone and interaction design, AI constraints, and human-centered design for LLM-powered systems.

The emphasis is on designing how AI behaves in context, rather than on technical novelty or implementation details.

Closing

If you’re curious about conversational AI, multimodal experiences, or how AI agents can be designed responsibly, feel free to explore the assistant or share your thoughts.

I’m always open to conversations about interesting product and AI challenges.