Logo
Your Comeback Starts Here
Mission First
Dynamic Roadmap Generation
Client-Side PDF Generation
Generated PDF Cover page
Real-Time Voice Practice
Actionable Feedback

Inspiration

The tech industry moves at breakneck speed. A two-year career break can feel like a decade in technology years. We were inspired by the countless talented women who stepped away from tech careers—for maternity leave, caregiving responsibilities, or personal reasons—only to find themselves facing an intimidating chasm when they tried to return.

It isn't just about the skills gap; it is about the confidence gap. We realized that generic tutorials aren't enough. What's missing is a personalized bridge: a tool that looks at where you were, understands where the industry is now, and holds your hand through the preparation process. We built Career Comeback Coach (CCC) to replace "overwhelm" with "actionable steps."

What it does

Career Comeback Coach is a privacy-first, AI-powered mentorship platform designed to fast-track a user's return to the IT industry.

Smart Resume Analysis: Users upload their old resume or manually input their past experience.

Tailored Roadmap Generation: Using Google Gemini, the system generates a custom week-by-week study plan (customizable from 2 to 12 weeks). This covers exactly what to revise and what new technologies to learn based on their specific background.

Professional PDF Export: Users can download their roadmap in a professionally designed PDF format. This serves as an offline guide for their weekly goals and resources.

AI Mock Interviews (Dual Mode):

Text-Based: A chat interface where the AI asks questions, and responses are read aloud using ElevenLabs TTS.
Voice-to-Voice: A fully conversational experience powered by ElevenLabs Agents integrated with Google Gemini. The user speaks naturally, and the AI interviewer responds instantly with human-like audio, creating an immersive simulation of a real recruiter interaction.

"Resume Your Journey": Since we don't store data on a server, we built a feature where users can upload their previously generated PDF roadmap. The AI reads the file and tailors a new mock interview session based on that specific plan, allowing users to practice multiple times without regenerating the roadmap.

How we built it

We prioritized a lightweight, Serverless, Client-Side First architecture to ensure speed and absolute privacy.

Frontend Core: The entire application is built with Angular v20, utilizing the latest features (Signals, Standalone Components) for a reactive and smooth user experience.

Direct API Integration: We bypassed the need for a backend server. The frontend communicates directly with the AI services via API keys.

The Brain (AI): We integrated Google Gemini for the heavy lifting—parsing resumes, generating the structured JSON for the roadmaps, and acting as the interview logic engine.

The Voice: We integrated ElevenLabs Conversational AI (Agents) to power the voice-to-voice interview experience, providing natural conversation flow with built-in speech recognition and synthesis. For text-based interactions, we utilize ElevenLabs' TTS API to read responses aloud.

Session Management: The application relies entirely on browser session storage. No database is used. User data exists only while the tab is open, ensuring that sensitive resume data is never permanently stored on our end.

Challenges we ran into

Managing State without a DB: Since we opted for a zero-backend approach for privacy, we had to figure out how a user could "save" their progress. This led to the development of the PDF-to-Session feature, where the downloaded roadmap acts as a "save file" that can be re-uploaded to restore the context.

Voice Latency & Flow: Creating a natural "Voice-to-Voice" experience required careful orchestration between ElevenLabs Agents and Gemini. We optimized the integration to ensure the conversational AI agent had access to the right context from the user's roadmap while maintaining smooth, natural turn-taking in the conversation.

API Limits: Managing the token limits and API quotas (specifically for high-quality voice synthesis) required us to implement hard limits (1-minute voice sessions, 5-minute text sessions) for the live demo to ensure stability without crashing our quota.

Accomplishments that we're proud of

Genuine Empathy in AI: We managed to tune the system prompt so the AI doesn't just check code—it offers encouragement, mirroring the supportive tone of a real career coach.

Client-Side PDF Generation: We successfully implemented a system where the AI's JSON output is rendered into a beautiful, downloadable PDF entirely in the browser, within seconds.

The "Resume Interview" Logic: We successfully built the feature where the AI can parse our custom-generated PDF roadmap and context-switch back into "Interviewer Mode" based on that specific data.

What we learned

Angular v20 Capabilities: pushing the boundaries of the latest Angular version showed us how efficient modern frontend frameworks have become at handling complex logic that used to require a backend.

Multimodal AI is Transformative: Gemini's ability to process resumes and context effectively allowed us to skip building complex parsing algorithms.

The Importance of Conversational AI: We learned that truly natural voice interactions require more than just good text-to-speech. ElevenLabs Agents' conversational capabilities, combined with Gemini's contextual understanding, created an interview experience that felt remarkably human—far beyond what we could achieve with separate STT/TTS components.

What's next for Career Comeback Coach

Persistency: Adding an optional secure backend (Supabase or Firebase) for users who want to save their progress across devices.

Extended Voice Modes: Optimizing the ElevenLabs Agents integration to allow for longer, more complex interview sessions with better context retention across multiple conversation turns.

Community Hub: Creating a space for users to share their roadmaps and success stories to build a support network.