The problem it solves

• Fragmented Interview Tools
  Technical interviews often require multiple disconnected tools for scheduling, coding, video calls, and evaluation.

• Lack of Real-Time Collaboration
  Many platforms fail to provide smooth, low-latency collaborative coding with proper version tracking.

• Unfair or Incomplete Evaluation
  Manual scoring and limited analytics make it hard to accurately assess problem-solving skills and performance.

• Cheating & Integrity Issues
  Existing solutions lack strong, real-time anti-cheat mechanisms during remote interviews.

• Poor Interview Experience
  Candidates and interviewers face friction due to unstable setups, accessibility gaps, and delayed feedback.

CodeCollab solves these issues by offering a unified, real-time, and AI-assisted interview platform with built-in collaboration, monitoring, and intelligent evaluation.

Challenges we ran into

• WebRTC Video Call Integration
  Integrating real-time video calls using WebRTC within an already complex system was challenging, especially while ensuring stability, low latency, and smooth coexistence with live coding and monitoring features.

• Tool Calling with Local LLM Setup
  Implementing tool/function calling using LM Studio in a local environment was new and non-trivial. Managing model behavior, request routing, and reliable responses required multiple iterations and careful prompt design.

• Real-Time Code Synchronization
  Achieving seamless real-time code sync between interviewer and candidate was challenging. Using Tiptap for collaborative editing required precise state management to ensure consistency, performance, and conflict-free updates.