A Voice-First, Low-Latency AI Mock Interviewer
Experience structured, hyper-realistic mock interviews conducted entirely by voice. Bodhi listens, evaluates, and dynamically adapts to your answers in real-time.
Watch the end-to-end interview flow, from introduction to the final technical deep-dive.
Bodhi is built for ultra-low latency and stateful conversations. It is orchestrated via LangGraph, streams real-time voice using Sarvam AI, and perfectly handles edge cases through a custom context-memory pipeline.
Bodhi isn't just a chatbot; it's a fully integrated interview environment.
When Bodhi asks a coding or Data Structures/Algorithms (DSA) question, an interactive text editor dynamically opens, allowing you to write code while explaining your thought process naturally.
Our computer-vision pipeline monitors the candidate at 1 FPS. If a user looks away, opens another tab, or uses a phone, Bodhi's proctoring system flags the violation and updates the candidate's trust rating instantly.
Bodhi isn't scripted. It gracefully transitions between Intro, Technical, Behavioral, and DSA phases. If an answer is vague, the AI actively challenges the candidate with targeted follow-up questions.
At the end of the session, Bodhi generates a rigorous, multi-dimensional performance report. It grades candidates across four axes: Accuracy, Depth, Communication, and Confidence, culminating in an agentic hiring recommendation.
- Resume Parsing & Gap Analysis: Upload a resume and Job Description. Bodhi automatically finds the gaps and tailors the interview questions to probe specific weak points.
- RAG-Powered Context: Ingests Company and Role profiles to ground the AI's questioning in real-world, company-specific contexts.
- Phase Memory Compaction: As the interview progresses, Bodhi summarizes previous phases, allowing it to naturally reference earlier answers (e.g., "Earlier you mentioned caching, how would that apply here?").
- Real-Time Audio Streaming: Energy-gated Voice Activity Detection (VAD) coupled with auto-chunking ensures natural, hands-free conversation without ever pressing an "Enter" key.
Built to redefine technical and behavioral screening.