Taksh

Taksh is a multi-agentic SaaS platform that revolutionizes mock interview preparation. Users submit a job description, and our AI agents orchestrate a complete interview experience—from personalized question generation to detailed performance analysis.

How It Works

1. User Onboarding & Profile Generation

Users submit their resume and optionally a LinkedIn profile export (PDF). Our first agent analyzes these documents to build a comprehensive user profile, identifying key skills, experience levels, and career trajectory.

Users upload resume PDF → AI extracts skills, experience, and generates profile

2. Job Description Submission & Analysis

From the dashboard, users submit job descriptions for positions they’re targeting. A dedicated agent parses the JD, extracting required skills, seniority level, company culture cues, and technical requirements.

User dashboard showing submitted JDs with analysis status

3. Interview Session Creation

Based on the JD analysis and user profile, our system automatically creates a tailored interview session. Questions are generated to match the specific role requirements, difficulty level, and the user’s experience gaps.

4. Pre-Interview Lobby

Before entering the interview, users land in a lobby where they can review what to expect: interview topics, estimated duration, question types, and tips for success.

Users review interview details, topics, and expectations before joining

5. Live Interview with AI Agent

Users connect with our live interview agent in a real-time session. The agent asks contextual questions based on the JD analysis, adapts difficulty based on responses, and collects performance data throughout the interview.

Real-time interview interface with AI agent asking contextual questions

6. Comprehensive Report Generation

After the interview, our final agent synthesizes all collected data—response quality, technical accuracy, communication skills, and areas for improvement—into a detailed performance report with actionable feedback.

Tech Stack

• Frontend: Next.js with Server Components for optimized rendering
• AI Agents: Python-based live interview agent with real-time processing
• Analytics: PostHog for feature flag-based deployment and user analytics
• Infrastructure: Containerized microservices with Kubernetes orchestration
• Real-time: WebSocket-based live interview sessions

Key Features

• Resume & LinkedIn profile parsing for automatic user profiling
• JD analysis engine that extracts role-specific requirements
• Dynamic question generation tailored to user experience level
• Real-time adaptive interviewing with difficulty adjustment
• Comprehensive performance analytics and actionable feedback
• Multi-agent orchestration for seamless user experience
• Feature flag-based deployment for gradual rollouts

Sample Report