I let L from Death Note grill my software experience while optical sensors monitored my blood flow

This is a submission for the Presage SmartSpectra SDK Challenge

What I Built

"I am L. Candidate kun... you claim on your resume that you architected distributed clusters handling fifty million requests per second. Yet the exact millisecond I asked you about cache invalidation, your optical pulse surged by nineteen beats per minute and your micro vascular blood volume doubled. Tell me... why are your capillaries confessing before you even open your mouth?"

Technical interviews in 2026 are completely broken.

Candidate resumes are stuffed with artificial intelligence buzzwords, exaggerated claims of scaling global cloud infrastructure, and rehearsed algorithm solutions memorized the night before. Standard whiteboard interviews can no longer distinguish genuine architectural masters from confident prompt engineers.

RejectNote is a satirical technical interview roasting chamber powered by optical contactless photoplethysmography polygraph surveillance. Instead of a polite human resources screen, iconic anime personalities interrogate your resume claims in real time.

Currently in Season 1, the master detective L Lawliet from Death Note sits across the table as your chief inquisitor. As you attempt to defend your architectural decisions, the Presage SmartSpectra SDK silently turns your standard camera feed into an optical polygraph. It monitors involuntary sub surface facial blood flow, cardiac pulse surges, respiration shifts, and physiological stress arousal directly through ambient light reflected off your skin.

When your heart rate accelerates or breathing tightens during a tough technical question, L immediately pounces on the biological tell, calls out your bluff with authentic voice audio, triggers dramatic anime visual reactions, and logs an indelible entry into the Rejection Note.

Watching Your Bluffs Unravel

• 👁️ Optical rPPG Capillary Scan: Analyzes sub surface light absorption in facial micro vessels via the webcam using the Presage SmartSpectra SDK. No wearables or external hardware touched.
• 💓 Cardiac Surge Polygraph: Calculates resting pulse baseline in the first 15 seconds. If a candidate bluffs and pulse spikes by over 15 BPM, L immediately pounces.
• 🫁 Respiration Hyperventilation Monitor: Tracks chest and diaphragm breathing cycles. When technical panic sets in, rapid breathing trips the suspicion gauge.
• ⚡ Scouter Ki Level and Power Surge: Smoothly calculates resume credibility while Goku goes Super Saiyan, scaling scouter power readings from 1,050 to over 9,000 with ki aura visuals.
• 📜 The Rejection Note Autopsy: Generates an indelible cryptographic report card documenting every physiological spike, timestamped lie, and final verdict.

Demo

Direct link: https://youtu.be/7mf08mP4sQg

Video Highlights:

• 0:00 Candidate profile intake and resume parsing
• 0:25 Optical camera calibration and baseline signal lock
• 0:50 Interrogation by L with live biometric polygraph telemetry
• 1:35 Stress surges trigger dynamic anime roasting and tactical interrogation
• 2:20 Cryptographic Rejection Note and final verdict

Code

The complete source code is public on GitHub:
https://github.com/Supersaiyan101/Reject_Note

Quick Setup Instructions

1. Clone the repository:

git clone https://github.com/Supersaiyan101/Reject_Note.git
cd Reject_Note

1. Install dependencies:

npm install

1. Configure environment keys: Create a .env file in the project root:

VITE_PRESAGE_API_KEY=your_presage_api_key_here

1. Launch the local development server:

npm run dev

Visit http://localhost:5173 in any modern Chromium browser with webcam access.

How It Responds to Human Signals

RejectNote integrates the Presage SmartSpectra SDK to extract rich physiological signals directly from ambient light reflections on the candidate facial skin:

1. Optical Signal Extraction:
   The application captures high frequency video frames client side. SmartSpectra analyzes subtle color variations in red, green, and blue spectrums caused by pulsatile blood volume changes in the micro vascular bed.

2. Physiological Baseline Calibration:
   During the intake stage, the system conducts a fifteen second resting baseline capture. It computes baseline resting pulse rate (BPM), breathing rate (RPM), and sympathetic tone index.

3. Dynamic Stress Response Engine:
   As L presents challenging questions about concurrency, caching, and database transactions, the app continuously compares incoming telemetry against the baseline:

   • Calm State (Delta under 8 BPM): L remains analytical and observant.
   • Agitated State (Delta between 8 and 15 BPM): L raises his thumb to his lip, warning the candidate that micro tremors suggest deception.
   • Caught Red Handed (Delta over 15 BPM or respiration surge over 30 percent): L slams the desk, red alert flashing fills the screen, custom voice audio fires, and the interrogation suspicion gauge maxes out.

4. Multi Modal Audio and Visual Feedback:
   Biometric spikes directly manipulate the visual novel engine. L shifts between calm, suspicious, and aggressive postures, accompanied by dramatic anime sound effects, heart beat audio, and spoken deductions via the Web Speech API.

Live Biometric Interrogation Log

Here is the exact biometric telemetry recorded during our live interrogation run:

• Phase 1: Distributed Redis Caching

  • Candidate Claim: "Maintained 99.999% SLA across 4 cloud regions"
  • Biometric Telemetry: 68 BPM surging to 89 BPM (+21 BPM delta surge)
  • L Deduction: "Your pulse surged under 64 cameras. A classic sympathetic tell."
  • Verdict: 🚨 Caught Red Handed

• Phase 2: Kafka Event Streaming

  • Candidate Claim: "Zero message loss during network partition"
  • Biometric Telemetry: 14 RPM surging to 27 RPM (+92% breathing acceleration)
  • L Deduction: "Hyperventilation detected. You did not configure failover, did you?"
  • Verdict: 🚨 Bluff Exposed

• Phase 3: PostgreSQL Isolation Levels

  • Candidate Claim: "Explained serializable snapshot isolation calmly"
  • Biometric Telemetry: 70 BPM to 71 BPM (Stable resting baseline)
  • L Deduction: "Fascinating. Your capillaries remained steady. Suspicion reduced to 5%."
  • Verdict: ✅ Plausible Truth

Keeping It Honest

This was vital during development:

• Authentic Presage Optical Telemetry: The biometric signals are not pre scripted mock data. The webcam stream is fed directly into the Presage SmartSpectra SDK, extracting live physiological signals client side in real time.
• Deterministic Reaction Thresholds: L does not insult you randomly. Every audio cue, anime visual reaction, and screen shake is mathematically triggered by deltas exceeding your calibrated resting baseline.
• Zero Cloud Video Surveillance: All video processing occurs entirely in browser memory. No camera frames ever leave your personal computer.

Prize Categories

Most Creative or Delightful Use

RejectNote transforms a universally dreaded rite of passage (the intimidating technical interview) into a hilarious, gamified, and unforgettable interactive anime showdown.

Instead of treating biometric telemetry as a dry clinical readout, RejectNote repurposes the cutting edge contactless sensing of the Presage SmartSpectra SDK into an anime polygraph. Pairing L Lawliet's legendary analytical interrogation style with real sub surface blood flow detection creates an unexpected, playful, and delightfully tense experience where you cannot hide behind rehearsed answers because your capillaries will betray you.

The Spirit of It

The moment this project truly landed for me was during my first late night test run. I wrote the interrogation questions myself and knew exactly what L was going to say.

Yet the second L crouched forward, locked eyes with me, and asked how I handled production Redis failovers while my webcam feed zoomed in, my pulse spontaneously jumped by nineteen beats per minute on screen. My palms actually sweated. That was the moment I realized: you can rehearse prompt engineering, and you can memorize LeetCode solutions, but your own capillaries will never lie for you.

Agent Session

I built and architected RejectNote as a solo developer collaborating with an AI coding agent.

The creative vision, psychological anime interrogation concept, and software architecture were entirely mine. I used the AI assistant as a tireless pair programmer to help scaffold the React and TypeScript front end, write Web Audio synthesis algorithms, fine tune Tailwind styling, and integrate the Presage SmartSpectra SDK client library. Every interrogation branch, anime voice cue, biometric threshold, and test run was directed, inspected, and verified by me.

Consent and Data Handling

User privacy, security, and consent are built into every layer of RejectNote:

• Explicit Camera Consent: Camera access is requested strictly through browser permission dialogs with prominent visual indicators.
• Pure Client Side Processing: Video frames are analyzed entirely in browser memory. Video feeds are never recorded, never written to disk, and never transmitted to external cloud servers.
• Ephemeral Biometrics: Extracted physiological metrics (pulse and respiration) exist exclusively within local React application state for the duration of the interrogation session and are discarded on page refresh.
• Zero Persistent Facial Identification: RejectNote does not perform facial recognition, facial template storage, or identity tracking.
• Non Medical Entertainment Disclaimer: RejectNote is designed strictly for entertainment and educational purposes. It does not provide medical diagnoses, clinical evaluations, or physiological treatment advice.